Tootfinder

Dolphins QB Tua Tagovailoa says relationship with WR Tyreek Hill 'still a work in progress' https://www.nfl.com/news/dolphins-qb-tua-tagovailoa-says-relationship-with-wr-tyreek-hill-still-a-work-in-progress

Raiders Confident in Group That is a Work in Progress https://www.si.com/nfl/raiders/news/raiders-training-camp-elandon-roberts-devin-white-pete-carroll

AnyMAC: Cascading Flexible Multi-Agent Collaboration via Next-Agent Prediction
Song Wang, Zhen Tan, Zihan Chen, Shuang Zhou, Tianlong Chen, Jundong Li
https://arxiv.org/abs/2506.17784

AnyMAC: Cascading Flexible Multi-Agent Collaboration via Next-Agent Prediction
Recent progress in large language model (LLM)-based multi-agent collaboration highlights the power of structured communication in enabling collective intelligence. However, existing methods largely rely on static or graph-based inter-agent topologies, lacking the potential adaptability and flexibility in communication. In this work, we propose a new framework that rethinks multi-agent coordination through a sequential structure rather than a graph structure, offering a significantly larger topo…

Historical Tech Tree
The tech tree is an interactive visualization of technological history from 3 million years ago to today. A work in progress, it currently contains 1988 technologies and 2369 connections between them.
⏳ https://www.historicaltechtree.com

Historical Tech Tree
Interactive visualization of technological history

Understanding the Drag Torque in Common Envelope Evolution
Soumik Bhattacharyya, Luke Chamandy, Eric G. Blackman, Adam Frank, Baowei Liu
https://arxiv.org/abs/2506.19547

Understanding the Drag Torque in Common Envelope Evolution
Common envelope (CE) evolution is largely governed by the drag torque applied on the inspiralling stellar components by the envelope. Previous work has shown that idealized models of the torque based on a single body moving in rectilinear motion through an unperturbed atmosphere can be highly inaccurate. Progress requires new models for the torque that account for binarity. Toward this end we perform a new 3D global hydrodynamic CE simulation with the mass of the companion point particle set eq…

OpenNav: Open-World Navigation with Multimodal Large Language Models
Mingfeng Yuan, Letian Wang, Steven L. Waslander
https://arxiv.org/abs/2507.18033 https://

OpenNav: Open-World Navigation with Multimodal Large Language Models
Pre-trained large language models (LLMs) have demonstrated strong common-sense reasoning abilities, making them promising for robotic navigation and planning tasks. However, despite recent progress, bridging the gap between language descriptions and actual robot actions in the open-world, beyond merely invoking limited predefined motion primitives, remains an open challenge. In this work, we aim to enable robots to interpret and decompose complex language instructions, ultimately synthesizing a…

Better Bounds for Semi-Streaming Single-Source Shortest Paths
Sepehr Assadi, Gary Hoppenworth, Janani Sundaresan
https://arxiv.org/abs/2507.17841 https://a…

Better Bounds for Semi-Streaming Single-Source Shortest Paths
In the semi-streaming model, an algorithm must process any $n$-vertex graph by making one or few passes over a stream of its edges, use $O(n \cdot \text{polylog }n)$ words of space, and at the end of the last pass, output a solution to the problem at hand. Approximating (single-source) shortest paths on undirected graphs is a longstanding open question in this model. In this work, we make progress on this question from both upper and lower bound fronts: We present a simple randomized algorith…

Spotted this Progress Pride Graffiti at the weekend in a carpark that I often use. There's some awesome work there!
#Graffiti #ProgressPride #Pride

The results of the development of the SPHERE-3 detector for studying the PCR mass composition in the 1-1000 PeV energy range. The status of 2025
D. V. Chernov, E. A. Bonvech, O. V. Cherkesova, E. L. Entina, V. I. Galkin, V. A. Ivanov, T. A. Kolodkin, N. O. Ovcharenko, D. A. Podgrudkov, T. M. Roganova, M. D. Ziva
https://arxiv.or…

The results of the development of the SPHERE-3 detector for studying the PCR mass composition in the 1-1000 PeV energy range. The status of 2025
The SPHERE-3 setup is destined to register the EAS Cherenkov light, both the direct and the reflected from the snow surface. A set of methods and new approaches in the measurement technique enable substantial progress in the study of the primary cosmic ray composition in the energy range 1--1000 PeV. The present work reveals the current status of the project development and detector performance modelling.

Lower Bounds against the Ideal Proof System in Finite Fields
Tal Elbaz, Nashlen Govindasamy, Jiaqi Lu, Iddo Tzameret
https://arxiv.org/abs/2506.17210 https…

Lower Bounds against the Ideal Proof System in Finite Fields
Lower bounds against strong algebraic proof systems and specifically fragments of the Ideal Proof System (IPS), have been obtained in an ongoing line of work. All of these bounds, however, are proved only over large (or characteristic $0$) fields, yet finite fields are the more natural setting for propositional proof complexity, especially for progress toward lower bounds for Frege systems such as $AC^0[p]$-Frege. This work establishes lower bounds against fragments of IPS over fixed finite fie…

Various thi.ng updates, bug fixes, additions and new version of https://github.com/thi-ng/zig-thing/ — now fully compatible with current Zig v0.14.1
On a more diary/devlog note: I also updated several of my Zig based work-in-progress art pieces to the latest version (some of them not touc…

Should we teach vibe coding? Here's why not.
Should AI coding be taught in undergrad CS education?
1/2
I teach undergraduate computer science labs, including for intro and more-advanced core courses. I don't publish (non-negligible) scholarly work in the area, but I've got years of craft expertise in course design, and I do follow the academic literature to some degree. In other words, In not the world's leading expert, but I have spent a lot of time thinking about course design, and consider myself competent at it, with plenty of direct experience in what knowledge & skills I can expect from students as they move through the curriculum.
I'm also strongly against most uses of what's called "AI" these days (specifically, generative deep neutral networks as supplied by our current cadre of techbro). There are a surprising number of completely orthogonal reasons to oppose the use of these systems, and a very limited number of reasonable exceptions (overcoming accessibility barriers is an example). On the grounds of environmental and digital-commons-pollution costs alone, using specifically the largest/newest models is unethical in most cases.
But as any good teacher should, I constantly question these evaluations, because I worry about the impact on my students should I eschew teaching relevant tech for bad reasons (and even for his reasons). I also want to make my reasoning clear to students, who should absolutely question me on this. That inspired me to ask a simple question: ignoring for one moment the ethical objections (which we shouldn't, of course; they're very stark), at what level in the CS major could I expect to teach a course about programming with AI assistance, and expect students to succeed at a more technically demanding final project than a course at the same level where students were banned from using AI? In other words, at what level would I expect students to actually benefit from AI coding "assistance?"
To be clear, I'm assuming that students aren't using AI in other aspects of coursework: the topic of using AI to "help you study" is a separate one (TL;DR it's gross value is not negative, but it's mostly not worth the harm to your metacognitive abilities, which AI-induced changes to the digital commons are making more important than ever).
So what's my answer to this question?
If I'm being incredibly optimistic, senior year. Slightly less optimistic, second year of a masters program. Realistic? Maybe never.
The interesting bit for you-the-reader is: why is this my answer? (Especially given that students would probably self-report significant gains at lower levels.) To start with, [this paper where experienced developers thought that AI assistance sped up their work on real tasks when in fact it slowed it down] (https://arxiv.org/abs/2507.09089) is informative. There are a lot of differences in task between experienced devs solving real bugs and students working on a class project, but it's important to understand that we shouldn't have a baseline expectation that AI coding "assistants" will speed things up in the best of circumstances, and we shouldn't trust self-reports of productivity (or the AI hype machine in general).
Now we might imagine that coding assistants will be better at helping with a student project than at helping with fixing bugs in open-source software, since it's a much easier task. For many programming assignments that have a fixed answer, we know that many AI assistants can just spit out a solution based on prompting them with the problem description (there's another elephant in the room here to do with learning outcomes regardless of project success, but we'll ignore this over too, my focus here is on project complexity reach, not learning outcomes). My question is about more open-ended projects, not assignments with an expected answer. Here's a second study (by one of my colleagues) about novices using AI assistance for programming tasks. It showcases how difficult it is to use AI tools well, and some of these stumbling blocks that novices in particular face.
But what about intermediate students? Might there be some level where the AI is helpful because the task is still relatively simple and the students are good enough to handle it? The problem with this is that as task complexity increases, so does the likelihood of the AI generating (or copying) code that uses more complex constructs which a student doesn't understand. Let's say I have second year students writing interactive websites with JavaScript. Without a lot of care that those students don't know how to deploy, the AI is likely to suggest code that depends on several different frameworks, from React to JQuery, without actually setting up or including those frameworks, and of course three students would be way out of their depth trying to do that. This is a general problem: each programming class carefully limits the specific code frameworks and constructs it expects students to know based on the material it covers. There is no feasible way to limit an AI assistant to a fixed set of constructs or frameworks, using current designs. There are alternate designs where this would be possible (like AI search through adaptation from a controlled library of snippets) but those would be entirely different tools.
So what happens on a sizeable class project where the AI has dropped in buggy code, especially if it uses code constructs the students don't understand? Best case, they understand that they don't understand and re-prompt, or ask for help from an instructor or TA quickly who helps them get rid of the stuff they don't understand and re-prompt or manually add stuff they do. Average case: they waste several hours and/or sweep the bugs partly under the rug, resulting in a project with significant defects. Students in their second and even third years of a CS major still have a lot to learn about debugging, and usually have significant gaps in their knowledge of even their most comfortable programming language. I do think regardless of AI we as teachers need to get better at teaching debugging skills, but the knowledge gaps are inevitable because there's just too much to know. In Python, for example, the LLM is going to spit out yields, async functions, try/finally, maybe even something like a while/else, or with recent training data, the walrus operator. I can't expect even a fraction of 3rd year students who have worked with Python since their first year to know about all these things, and based on how students approach projects where they have studied all the relevant constructs but have forgotten some, I'm not optimistic seeing these things will magically become learning opportunities. Student projects are better off working with a limited subset of full programming languages that the students have actually learned, and using AI coding assistants as currently designed makes this impossible. Beyond that, even when the "assistant" just introduces bugs using syntax the students understand, even through their 4th year many students struggle to understand the operation of moderately complex code they've written themselves, let alone written by someone else. Having access to an AI that will confidently offer incorrect explanations for bugs will make this worse.
To be sure a small minority of students will be able to overcome these problems, but that minority is the group that has a good grasp of the fundamentals and has broadened their knowledge through self-study, which earlier AI-reliant classes would make less likely to happen. In any case, I care about the average student, since we already have plenty of stuff about our institutions that makes life easier for a favored few while being worse for the average student (note that our construction of that favored few as the "good" students is a large part of this problem).
To summarize: because AI assistants introduce excess code complexity and difficult-to-debug bugs, they'll slow down rather than speed up project progress for the average student on moderately complex projects. On a fixed deadline, they'll result in worse projects, or necessitate less ambitious project scoping to ensure adequate completion, and I expect this remains broadly true through 4-6 years of study in most programs (don't take this as an endorsement of AI "assistants" for masters students; we've ignored a lot of other problems along the way).
There's a related problem: solving open-ended project assignments well ultimately depends on deeply understanding the problem, and AI "assistants" allow students to put a lot of code in their file without spending much time thinking about the problem or building an understanding of it. This is awful for learning outcomes, but also bad for project success. Getting students to see the value of thinking deeply about a problem is a thorny pedagogical puzzle at the best of times, and allowing the use of AI "assistants" makes the problem much much worse. This is another area I hope to see (or even drive) pedagogical improvement in, for what it's worth.
1/2

Graph-Coarsening for Machine Learning Coarse-grained Molecular Dynamics
Soumya Mondal, Subhanu Halder, Debarchan Basu, Sandeep Kumar, Tarak Karmakar
https://arxiv.org/abs/2507.16531

Graph-Coarsening for Machine Learning Coarse-grained Molecular Dynamics
Coarse-grained (CG) molecular dynamics (MD) simulations can simulate large molecular complexes over extended timescales by reducing degrees of freedom. A critical step in CG modeling is the selection of the CG mapping algorithm, which directly influences both accuracy and interpretability of the model. Despite progress, the optimal strategy for coarse-graining remains a challenging task, highlighting the necessity for a comprehensive theoretical framework. In this work, we present a graph-based…

Raiders’ long ball connection remains a work in progress https://www.reviewjournal.com/sports/raiders/raiders-long-ball-connection-remains-a-work-in-progress-3422923/

WIP: Leveraging LLMs for Enforcing Design Principles in Student Code: Analysis of Prompting Strategies and RAG
Dhruv Kolhatkar, Soubhagya Akkena, Edward F. Gehringer
https://arxiv.org/abs/2508.11717

WIP: Leveraging LLMs for Enforcing Design Principles in Student Code: Analysis of Prompting Strategies and RAG
This work-in-progress research-to-practice paper explores the integration of Large Language Models (LLMs) into the code-review process for open-source software projects developed in computer science and software engineering courses. The focus is on developing an automated feedback tool that evaluates student code for adherence to key object-oriented design principles, addressing the need for more effective and scalable methods to teach software design best practices. The innovative practice inv…

Did a little 9 mile loop this morning on the bike... Later on the way to work I spotted the progress of the new bike path on Lisbon Avenue. Lisbon has gone from a street I didn't even like to drive on to a street I look forward to biking on.
#mke #milwaukee

From Near-Integrable to Far-from-Integrable: A Unified Picture of Thermalization and Heat Transport
Weicheng Fu, Zhen Wang, Yisen Wang, Yong Zhang, Hong Zhao
https://arxiv.org/abs/2508.15566

From Near-Integrable to Far-from-Integrable: A Unified Picture of Thermalization and Heat Transport
Whether and how a system approaches equilibrium is central in nonequilibrium statistical physics, crucial to understanding thermalization and transport. Bogoliubov's three-stage (initial, kinetic, and hydrodynamic) evolution hypothesis offers a qualitative framework, but quantitative progress has focused on near-integrable systems like dilute gases. In this work, we investigate the relaxation dynamics of a one-dimensional diatomic hard-point (DHP) gas, presenting a phase diagram that characteri…

@… Do share screenshots, etc. if you’re up for it. I’d love to see your work in progress. 😄

ToxiFrench: Benchmarking and Enhancing Language Models via CoT Fine-Tuning for French Toxicity Detection
Axel Delaval, Shujian Yang, Haicheng Wang, Han Qiu, Jialiang Lu
https://arxiv.org/abs/2508.11281

ToxiFrench: Benchmarking and Enhancing Language Models via CoT Fine-Tuning for French Toxicity Detection
Detecting toxic content using language models is crucial yet challenging. While substantial progress has been made in English, toxicity detection in French remains underdeveloped, primarily due to the lack of culturally relevant, large-scale datasets. In this work, we introduce TOXIFRENCH, a new public benchmark of 53,622 French online comments, constructed via a semi-automated annotation pipeline that reduces manual labeling to only 10% through high-confidence LLM-based pre-annotation and huma…

DMOSpeech 2: Reinforcement Learning for Duration Prediction in Metric-Optimized Speech Synthesis
Yinghao Aaron Li, Xilin Jiang, Fei Tao, Cheng Niu, Kaifeng Xu, Juntong Song, Nima Mesgarani
https://arxiv.org/abs/2507.14988

DMOSpeech 2: Reinforcement Learning for Duration Prediction in Metric-Optimized Speech Synthesis
Diffusion-based text-to-speech (TTS) systems have made remarkable progress in zero-shot speech synthesis, yet optimizing all components for perceptual metrics remains challenging. Prior work with DMOSpeech demonstrated direct metric optimization for speech generation components, but duration prediction remained unoptimized. This paper presents DMOSpeech 2, which extends metric optimization to the duration predictor through a reinforcement learning approach. The proposed system implements a nove…

How popular media gets love wrong
Now a bit of background about why I have this "engineered" model of love:
First, I'm a white straight cis man. I've got a few traits that might work against my relationship chances (e.g., neurodivergence; I generally fit pretty well into the "weird geek" stereotype), but as I was recently reminded, it's possible my experience derives more from luck than other factors, and since things are tilted more in my favor than most people on the planet, my advice could be worse than useless if it leads people towards strategies that would only have worked for someone like me. I don't *think* that's the case, but it's worth mentioning explicitly.
When I first started dating my now-wife, we were both in graduate school. I was 26, and had exactly zero dating/romantic experience though that point in my life. In other words, a pretty stereotypical "incel" although I definitely didn't subscribe to incel ideology at all. I felt lonely, and vaguely wanted a romantic relationship (I'm neither aromantic nor asexual), but had never felt socially comfortable enough to pursue one before. I don't drink and dislike most social gatherings like parties or bars; I mostly hung around the fringes of the few college parties I attended, and although I had a reasonable college social life in terms of friends, I didn't really do anything to pursue romance, feeling too awkward to know where to start. I had the beginnings of crushes in both high school and college, but never developed a really strong crush, probably correlated with not putting myself in many social situations outside of close all-male friend gatherings. I never felt remotely comfortable enough to act on any of the proto-crushes I did have. I did watch porn and masturbate, so one motivation for pursuing a relationship was physical intimacy, but loneliness was as much of a motivating factor, and of course the social pressure to date was a factor too, even though I'm quite contrarian.
When I first started dating my now-wife, we were both in graduate school. I was 26, and had exactly zero dating/romantic experience though that point in my life. In other words, a pretty stereotypical "incel" although I definitely didn't subscribe to incel ideology at all. I felt lonely, and vaguely wanted a romantic relationship (I'm neither aromantic nor asexual), but had never felt socially comfortable enough to pursue one before. I don't drink and dislike most social gatherings like parties or bars; I mostly hung around the fringes of the few college parties I attended, and although I had a reasonable college social life in terms of friends, I didn't really do anything to pursue romance, feeling too awkward to know where to start. I had the beginnings of crushes in both high school and college, but never developed a really strong crush, probably correlated with not putting myself in many social situations outside of close all-male friend gatherings. I never felt remotely comfortable enough to act on any of the proto-crushes I did have. I did watch porn and masturbate, so one motivation for pursuing a relationship was physical intimacy, but loneliness was as much of a motivating factor, and of course the social pressure to date was a factor too, even though I'm quite contrarian.
I'm lucky in that I had some mixed-gender social circles already like intramural soccer and a graduate-student housing potluck. Graduate school makes a *lot* more of these social spaces accessible, so I recognize that those not in school of some sort have a harder time of things, especially if like me they don't feel like they fit in in typical adult social spaces like bars.
However, at one point I just decided that my desire for a relationship would need action on my part and so I'd try to build a relationship and see what happened. I worked up my courage and asked one of the people in my potluck if she'd like to go for a hike (pretty much clearly a date but not explicitly one; in retrospect not the best first-date modality in a lot of ways, but it made a little more sense in our setting where we could go for a hike from our front door). To emphasize this point: I was not in love with (or even infatuated with) my now-wife at that point. I made a decision to be open to building a relationship, but didn't follow the typical romance story formula beyond that. Now of course, in real life as opposed to popular media, this isn't anything special. People ask each other out all the time just because they're lonely, and some of those relationships turn out fine (although many do not).
I was lucky in that some aspects of who I am and what I do happened to be naturally comforting to my wife (natural advantage in the "appeal" model of love) but of course there are some aspects of me that annoy my wife, and we negotiate that. In the other direction, there's some things I instantly liked about my wife, and other things that still annoy me. We've figured out how to accept a little, change a little, and overall be happy with each other (though we do still have arguments; it's not like the operation/construction/maintenance of the "love mechanism" is always perfectly smooth). In particular though, I approached the relationship with the attitude of "I want to try to build a relationship with this person," at first just because of my own desires for *any* relationship, and then gradually more and more through my desire to build *this specific* relationship as I enjoyed the rewards of companionship.
So for example, while I think my wife is objectively beautiful, she's also *subjectively* very beautiful *to me* because having decided to build a relationship with her, I actively tried to see her as beautiful, rather than trying to judge whether I wanted a relationship with her based on her beauty. In other words, our relationship is more causative of her beauty-to-me than her beauty-to-me is causative of our relationship. This is the biggest way I think the "engineered" model of love differs from the "fire" and "appeal" models: you can just decide to build love independent of factors we typically think of as engendering love (NOT independent of your partner's willingness to participate, of course), and then all of those things like "thinking your partner is beautiful" can be a result of the relationship you're building. For sure those factors might affect who is willing to try building a relationship with you in the first place, but if more people were willing to jump into relationship building (not necessarily with full commitment from the start) without worrying about those other factors, they might find that those factors can come out of the relationship instead of being prerequisites for it. I think this is the biggest failure of the "appeal" model in particular: yes you *do* need to do things that appeal to your partner, but it's not just "make myself lovable" it's also: is your partner putting in the effort to see the ways that you are beautiful/lovable/etc., or are they just expecting you to become exactly some perfect person they've imagined (and/or been told to desire by society)? The former is perfectly possible, and no less satisfying than the latter.
To cut off my rambling a bit here, I'll just add that in our progress from dating through marriage through staying-married, my wife and I have both talked at times explicitly about commitment, and especially when deciding to get married, I told her that I knew I couldn't live up to the perfect model of a husband that I'd want to be, but that if she wanted to deepen our commitment, I was happy to do that, and so we did. I also rearranged my priorities at that point, deciding that I knew I wanted to prioritize this relationship above things like my career or my research interests, and while I've not always been perfect at that in my little decisions, I've been good at holding to that in my big decisions at least. In the end, *once we had built a somewhat-committed relationship*, we had something that we both recognized was worth more than most other things in life, and that let us commit even more, thus getting even more out of it in the long term. Obviously you can't start the first date with an expectation of life-long commitment, and you need to synchronize your increasing commitment to a relationship so that it doesn't become lopsided, which is hard. But if you take the commitment as an active decision and as the *precursor* to things like infatuation, attraction, etc., you can build up to something that's incredibly strong and rewarding.
I'll follow this up with one more post trying to distill some advice from my ramblings.
#relationships #love

Is Transfer Learning Necessary for Violin Transcription?
Yueh-Po Peng, Ting-Kang Wang, Li Su, Vincent K. M. Cheung
https://arxiv.org/abs/2508.13516 https://

Is Transfer Learning Necessary for Violin Transcription?
Automatic music transcription (AMT) has achieved remarkable progress for instruments such as the piano, largely due to the availability of large-scale, high-quality datasets. In contrast, violin AMT remains underexplored due to limited annotated data. A common approach is to fine-tune pretrained models for other downstream tasks, but the effectiveness of such transfer remains unclear in the presence of timbral and articulatory differences. In this work, we investigate whether training from scra…

Statistics in 3d gravity from knots and links
Jeevan Chandra
https://arxiv.org/abs/2508.10864 https://arxiv.org/pdf/2508.10864

Statistics in 3d gravity from knots and links
In recent years, there has been remarkable progress in evaluating wormhole amplitudes in 3d Einstein gravity with negative cosmological constant and matching them to statistics of 2d CFT data. In this work, we compute non-perturbative Gaussian and non-Gaussian gravitational contributions to the OPE statistics using a framework that can systematically generate a class of such non-perturbative effects - \textit{Fragmentation of knots and links by Wilson lines}. We illustrate this idea by construc…

Continuing on my Meshtastic kick, (probably because I keep buying radios to play with...) This time I have a configuration guide for common settings I've found which are useful. Still a work in progress but I think most of the common options are there.
If I've forgotten something, gotten something wrong, or you have a trick I should add, let me know!

Useful Meshtastic Config Options
This guide provides a breakdown of common settings I have found to be useful for general use cases on the Meshtastic network with standard devices. For a full breakdown of all options supported in Meshtastic nodes, refer to the official documentation for Meshtastic.

WIP: Exploring the Value of a Debugging Cheat Sheet and Mini Lecture in Improving Undergraduate Debugging Skills and Mindset
Andrew Ash, John Hu
https://arxiv.org/abs/2506.11339

WIP: Exploring the Value of a Debugging Cheat Sheet and Mini Lecture in Improving Undergraduate Debugging Skills and Mindset
This work-in-progress research paper explores the efficacy of a small-scale microelectronics debugging education intervention utilizing quasi-experimental design in an introductory microelectronics course for third-year electrical and computer engineering (ECE) students. In the first semester of research, the experimental group attended a debugging "mini lecture" covering two common sources of circuit error and received a debugging cheat sheet with recommendations for testing and hypothesis for…

WIP: Turning Fake Chips into Learning Opportunities
Haniye Mehraban, Saad Azmeen-ur-Rahman, John Hu
https://arxiv.org/abs/2507.13281 https://

WIP: Turning Fake Chips into Learning Opportunities
This work-in-progress paper presents a case study in which counterfeit TL074 operational amplifiers, discovered in a junior level electronics course, became the basis for a hands on learning experience. Counterfeit integrated circuits (IC) are increasingly common, posing a significant threat to the integrity of undergraduate electronics laboratories. Instead of simply replacing the counterfeit components, we turned the issue into a teaching moment. Students engaged in hands-on diagnostics measu…

Vela: Scalable Embeddings with Voice Large Language Models for Multimodal Retrieval
Ruofan Hu, Yan Xia, Minjie Hong, Jieming Zhu, Bo Chen, Xiaoda Yang, Minghui Fang, Tao Jin
https://arxiv.org/abs/2506.14445

Vela: Scalable Embeddings with Voice Large Language Models for Multimodal Retrieval
Multimodal large language models (MLLMs) have seen substantial progress in recent years. However, their ability to represent multimodal information in the acoustic domain remains underexplored. In this work, we introduce Vela, a novel framework designed to adapt MLLMs for the generation of universal multimodal embeddings. By leveraging MLLMs with specially crafted prompts and selected in-context learning examples, Vela effectively bridges the modality gap across various modalities. We then prop…

Work in progress
It's almost time! We can't wait to welcome you to Berlin Buzzwords tomorrow for our 16th edition!
Grab your Last Minute Tickets now: https://tickets.plainschwarz.com/bbuzz25/c/7bjWaoGyO/

Tickets
Join us for two days of exciting talks, workshops and discussions, and of course our barcamp on Sunday! Berlin Buzzwords will take place as a hybrid conference online and in our venue Kulturbrauerei in Berlin.

Versatile Wavelength-Division Multiplexed Quantum Key Distribution Network Operating Simultaneously in the O and C Bands
Davide Scalcon, Matteo Padovan, Paolo Villoresi, Giuseppe Vallone, Marco Avesani
https://arxiv.org/abs/2507.11175

Versatile Wavelength-Division Multiplexed Quantum Key Distribution Network Operating Simultaneously in the O and C Bands
Ongoing technological progress is accelerating the commercial and global-scale deployment of Quantum Key Distribution (QKD). Its ability to enable unconditionally secure communication is expected to be a key feature of future telecommunication networks, and practical demonstrations of QKD network implementations in real-world environments are crucial for ensuring reliable adoption. In this work, we demonstrate a four-node photonic QKD network that employs versatile and cost-effective wavelength…

Clearly a work in progress. Lots of promise, but very disjointed while chemistry is built, fitness is gained, and adaptations to a faster and more physical league are happening. It's a lot of roster changes.
The defense look shaky, but often they were put in bad positions by turnovers with too many players committed. Outside backs and midfielders.
Thrilled for Chiesa to get one. Happy for Mo to continue the opening day streak.
Ekitike is my MOTM. He already looks comfor…

Scaling Learned Image Compression Models up to 1 Billion
Yuqi Li, Haotian Zhang, Li Li, Dong Liu, Feng Wu
https://arxiv.org/abs/2508.09075 https://arxiv.or…

Scaling Learned Image Compression Models up to 1 Billion
Recent advances in large language models (LLMs) highlight a strong connection between intelligence and compression. Learned image compression, a fundamental task in modern data compression, has made significant progress in recent years. However, current models remain limited in scale, restricting their representation capacity, and how scaling model size influences compression performance remains unexplored. In this work, we present a pioneering study on scaling up learned image compression mode…

Also, in other other #ThingUmbrella related news. My current consulting/advisor contract finishes end of the month, so I will have some time to write about various other recent updates (and unpublished work-in-progress) in early August... Please stay tuned! ✌️

🤩 The latest Safari 26 Beta does not disappoint in terms of #CSS features:
⚓️ Anchor Positioning!
🪄 Scroll-driven animations!
🌯 text-wrap: pretty!
🌈 contrast-color()!
⏲️ progress()!
and more!
And this sentence: “We value the principles of progressive enhancement and separation of concerns.” 😉
Great work, @…

News from WWDC25: WebKit in Safari 26 beta
Welcome to WWDC25!

#TooLong4FessHole I have been trying to become better friends with the partner of my bestie for over a decade since they hooked up - for the obvious reasons. I never felt as though I was making any progress; all of the work was being done by me and I still felt like I was receiving "disapproving father in law" vibes.
Recently I behaved poorly about something; I immediately recon…

Shipwright: Proving liveness of distributed systems with Byzantine participants
Derek Leung, Nickolai Zeldovich, Frans Kaashoek
https://arxiv.org/abs/2507.14080

Shipwright: Proving liveness of distributed systems with Byzantine participants
Ensuring liveness in a decentralized system, such as PBFT, is critical, because there may not be any single administrator that can restart the system if it encounters a liveness bug. At the same time, liveness is challenging to achieve because any single participant could be malicious, and yet the overall system must make forward progress. While verification is a promising approach for ensuring the absence of bugs, no prior work has been able to verify liveness for an executable implementation …

Integrated microheater on the 4H-silicon-carbide-on-insulator platform and its applications
Wenhan Sun, Ruixuan Wang, Jingwei Li, Haipeng Zhang, Zhensheng Jia, Qing Li
https://arxiv.org/abs/2506.15035

Integrated microheater on the 4H-silicon-carbide-on-insulator platform and its applications
Recent progress in the 4H-silicon-carbide-on-insulator (4H-SiCOI) platform has resulted in the demonstration of essential building blocks such as low-loss waveguides and microresonators. In this work, we add tunability to the 4H-SiCOI platform by integrating microheaters with compact microresonators. The strong thermo-optic effect in SiC enables a resonance tuning rate of $11.7$ pm/mW for a 36-$μ$m-radius SiC microring, with a maximum wavelength shift up to $2.4$ nm (300 GHz). The thermal time…

No More Marching: Learning Humanoid Locomotion for Short-Range SE(2) Targets
Pranay Dugar, Mohitvishnu S. Gadde, Jonah Siekmann, Yesh Godse, Aayam Shrestha, Alan Fern
https://arxiv.org/abs/2508.14098

No More Marching: Learning Humanoid Locomotion for Short-Range SE(2) Targets
Humanoids operating in real-world workspaces must frequently execute task-driven, short-range movements to SE(2) target poses. To be practical, these transitions must be fast, robust, and energy efficient. While learning-based locomotion has made significant progress, most existing methods optimize for velocity-tracking rather than direct pose reaching, resulting in inefficient, marching-style behavior when applied to short-range tasks. In this work, we develop a reinforcement learning approach…

Global linear drift-wave eigenmode structures on flux surfaces in stellarators: ion temperature gradient mode
Hongxuan Zhu, H. Chen, Z. Lin, A. Bhattacharjee
https://arxiv.org/abs/2506.12948

Global linear drift-wave eigenmode structures on flux surfaces in stellarators: ion temperature gradient mode
Turbulent transport greatly impacts the performance of stellarator magnetic confinement devices. While significant progress has been made on the numerical front, theoretical understanding of turbulence in stellarators is still lacking. In particular, due to nonaxisymmetry, different field lines couple within flux surfaces, the effects from which have yet to be adequately studied. In this work, we numerically simulate the linear electrostatic ion-temperature-gradient modes in stellarators using …

Proceedings 14th International Workshop on Trends in Functional Programming in Education
Rose Bohrer (AIST, Tokyo, JP)
https://arxiv.org/abs/2508.02305 https://

Proceedings 14th International Workshop on Trends in Functional Programming in Education
The goal of TFPIE is to gather researchers, teachers and professionals that use, or are interested in the use of, functional programming in education. TFPIE aims to be a venue where novel ideas, classroom-tested ideas and work-in-progress on the use of functional programming in education are discussed. The one-day workshop will foster a spirit of open discussion by having a review process for publication after the workshop.

Cowboys Coach Speaks Out On New QB’s Progress https://heavy.com/sports/nfl/dallas-cowboys/joe-milton-doing-well-progress/?adt_ei=[email]

Slow progress the last couple days because I've been going into the office for work and the commute is eating up my time I'd otherwise spend on such things.
But now it's the weekend and the switch engine board is coming together nicely.
Still have 514 nets to route - mostly the supervisor, line card management buses, and power supply but also some other odds and ends like the FPGA JTAG and part of the SPI flash.
Also I have to get the tach/PWM signals from the m…

Beyond Formal Semantics for Capabilities and Skills: Model Context Protocol in Manufacturing
Luis Miguel Vieira da Silva, Aljosha K\"ocher, Felix Gehlhoff
https://arxiv.org/abs/2506.11180

Beyond Formal Semantics for Capabilities and Skills: Model Context Protocol in Manufacturing
Explicit modeling of capabilities and skills -- whether based on ontologies, Asset Administration Shells, or other technologies -- requires considerable manual effort and often results in representations that are not easily accessible to Large Language Models (LLMs). In this work-in-progress paper, we present an alternative approach based on the recently introduced Model Context Protocol (MCP). MCP allows systems to expose functionality through a standardized interface that is directly consumab…

Thanking the @… folks for the excellent work they do, and especially for their upcoming support for security certificates for IP addresses which is nothing short of revolutionary for the future of the (Small) Web.

Getting ready to issue IP address certificates
Happy to see the progress on this. Thought you might be interested in seeing how important this feature is going to be for the Small Web: It’s going to be invaluable for helping decentralise the web. Thank you for the work you do 🙂

Evaluating LLMs on Chinese Idiom Translation
Cai Yang, Yao Dou, David Heineman, Xiaofeng Wu, Wei Xu
https://arxiv.org/abs/2508.10421 https://arxiv.org/pdf/…

Evaluating LLMs on Chinese Idiom Translation
Idioms, whose figurative meanings usually differ from their literal interpretations, are common in everyday language, especially in Chinese, where they often contain historical references and follow specific structural patterns. Despite recent progress in machine translation with large language models, little is known about Chinese idiom translation. In this work, we introduce IdiomEval, a framework with a comprehensive error taxonomy for Chinese idiom translation. We annotate 900 translation p…

Towards Automatic Evaluation and High-Quality Pseudo-Parallel Dataset Construction for Audio Editing: A Human-in-the-Loop Method
Yuhang Jia, Hui Wang, Xin Nie, Yujie Guo, Lianru Gao, Yong Qin
https://arxiv.org/abs/2508.11966

Towards Automatic Evaluation and High-Quality Pseudo-Parallel Dataset Construction for Audio Editing: A Human-in-the-Loop Method
Audio editing aims to manipulate audio content based on textual descriptions, supporting tasks such as adding, removing, or replacing audio events. Despite recent progress, the lack of high-quality benchmark datasets and comprehensive evaluation metrics remains a major challenge for both assessing audio editing quality and improving the task itself. In this work, we propose a novel approach for audio editing task by incorporating expert knowledge into both the evaluation and dataset constructio…

2-dimensional TFTs via modular $\infty$-operads
Jan Steinebrunner
https://arxiv.org/abs/2506.22104 https://arxiv.org/pdf/2506.22104…

2-dimensional TFTs via modular $\infty$-operads
This lecture series is based on joint work in progress with Shaul Barkan, as well as work in progress of the author. The five sections of these notes correspond to the five lectures, but more details have been added. $2$-dimensional topological field theories ($2$D TFTs) valued in vector spaces are commutative Frobenius algebras. The goal of this lecture series is to generalize from the $1$-category of vector spaces to any symmetric monoidal $\infty$-category $\mathcal{C}$, i.e. to study symm…

This https://arxiv.org/abs/2506.05678 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csLG_…

Numerical Investigation of Sequence Modeling Theory using Controllable Memory Functions
The evolution of sequence modeling architectures, from recurrent neural networks and convolutional models to Transformers and structured state-space models, reflects ongoing efforts to address the diverse temporal dependencies inherent in sequential data. Despite this progress, systematically characterizing the strengths and limitations of these architectures remains a fundamental challenge.In this work, we propose a synthetic benchmarking framework to evaluate how effectively different sequenc…

#PhantastikPrompts 10.06.:
Wie hältst du es mit geschlechtergerechter Sprache?
Wer bei mir mitliest, dürfte gesehen haben, dass ich nach Möglichkeit neutrale Formen verwende (und wo es sich absolut nicht ausgeht damit, den Stern verwende). Bis hin dazu, dass ich sowohl für den Eigengebrauch, als auch für den Fremdgebrauch und die Romanwelt ein paar Neologismen geschaffen ha…

MMHU: A Massive-Scale Multimodal Benchmark for Human Behavior Understanding
Renjie Li, Ruijie Ye, Mingyang Wu, Hao Frank Yang, Zhiwen Fan, Hezhen Hu, Zhengzhong Tu
https://arxiv.org/abs/2507.12463

MMHU: A Massive-Scale Multimodal Benchmark for Human Behavior Understanding
Humans are integral components of the transportation ecosystem, and understanding their behaviors is crucial to facilitating the development of safe driving systems. Although recent progress has explored various aspects of human behavior$\unicode{x2014}$such as motion, trajectories, and intention$\unicode{x2014}$a comprehensive benchmark for evaluating human behavior understanding in autonomous driving remains unavailable. In this work, we propose $\textbf{MMHU}$, a large-scale benchmark for hu…

This https://arxiv.org/abs/2405.08719 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_sta…

Addressing Misspecification in Simulation-based Inference through Data-driven Calibration
Driven by steady progress in deep generative modeling, simulation-based inference (SBI) has emerged as the workhorse for inferring the parameters of stochastic simulators. However, recent work has demonstrated that model misspecification can compromise the reliability of SBI, preventing its adoption in important applications where only misspecified simulators are available. This work introduces robust posterior estimation~(RoPE), a framework that overcomes model misspecification with a small re…

CLASS_SZ II: Notes and Examples of Fast and Accurate Calculations of Halo Model, Large Scale Structure and Cosmic Microwave Background Observables
Boris Bolliet, Aleksandra Kusiak, Fiona McCarthy, Alina Sabyr, Kristen Surrao, Jens Chluba, Carmen Embil Villagra, Simone Ferraro, Boryana Hadzhiyska, Dongwon Han, J. Colin Hill, Juan Francisco Mac\'ias-P\'erez, Mathew Madhavacheril, Abhishek Maniyar, Yogesh Mehta, Shivam Pandey, Emmanuel Schaan, Blake Sherwin, Alessio Spurio Mancini…

CLASS_SZ II: Notes and Examples of Fast and Accurate Calculations of Halo Model, Large Scale Structure and Cosmic Microwave Background Observables
These notes are very much work-in-progress and simply intended to showcase, in various degrees of details (and rigour), some of the cosmology calculations that class_sz can do. We describe the class_sz code in C, Python and Jax. Based on the Boltzmann code class, it can compute a wide range of observables relevant to current and forthcoming CMB and Large Scale Structure surveys. This includes galaxy shear and clustering, CMB lensing, thermal and kinetic Sunyaev and Zeldovich observables, Cosmic…

Reliable Magnetometry for Antiferromagnets and Thin Films: Correcting Substrate Artifacts in Mn3Sn/MgO Systems
Katarzyna Gas, Maciej Sawicki
https://arxiv.org/abs/2507.01385

Reliable Magnetometry for Antiferromagnets and Thin Films: Correcting Substrate Artifacts in Mn3Sn/MgO Systems
The rapid progress in antiferromagnetic and altermagnetic spintronics has led to increased interest in magnetic materials with vanishing net magnetization but strong spin-dependent transport properties. As thin films of such materials become central to device concepts, precise magnetic characterization is essential, for quantify intrinsic moments, and interpret transport signatures such as the anomalous Hall effect. In this work, we show that commercial MgO substrates, commonly used in epitaxia…

This https://arxiv.org/abs/2404.08578 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_mat…

Classification and nonexistence for $t$-structures on derived categories of schemes
This work establishes new results on the classification of $t$-structures for many subcategories of the derived category of quasi-coherent sheaves on a Noetherian scheme. Our work makes progress in two different directions. On one hand, we provide an improvement of a result of Takahashi on $t$-structures, generalizing it to the case of the bounded derived category of coherent sheaves on a quasi-compact CM-excellent scheme of finite Krull dimension. On the other hand, via independent techniques,…

Breaking a Logarithmic Barrier in the Stopping Time Convergence Rate of Stochastic First-order Methods
Yasong Feng, Yifan Jiang, Tianyu Wang, Zhiliang Ying
https://arxiv.org/abs/2506.23335

Breaking a Logarithmic Barrier in the Stopping Time Convergence Rate of Stochastic First-order Methods
This work provides a novel convergence analysis for stochastic optimization in terms of stopping times, addressing the practical reality that algorithms are often terminated adaptively based on observed progress. Unlike prior approaches, our analysis: 1. Directly characterizes convergence in terms of stopping times adapted to the underlying stochastic process. 2. Breaks a logarithmic barrier in existing results. Key to our results is the development of a Grönwall-type argument tailored to such…

AEGISS -- Atomic orbital and Entropy-based Guided Inference for Space Selection -- A novel semi-automated active space selection workflow for quantum chemistry and quantum computing applications
Fabio Tarocco, Pi A. B. Haase, Fabijan Pavo\v{s}evi\'c, Vijay Krishna, Leonardo Guidoni, Stefan Knecht, Martina Stella
https://arxiv.org/abs/2…

AEGISS -- Atomic orbital and Entropy-based Guided Inference for Space Selection -- A novel semi-automated active space selection workflow for quantum chemistry and quantum computing applications
The selection of a balanced active space is a critical step in multi-reference quantum chemistry calculations, particularly for systems with strong electron correlation. Likewise, active space selection is a key to unlock the potential of contemporary quantum computing in quantum chemistry. Albeit recent progress, there remains a lack of a unified, robust, and fully automated framework for active space selection that performs reliably across a wide range of molecular systems. In this work, we…

Promoting Efficient Reasoning with Verifiable Stepwise Reward
Chuhuai Yue, Chengqi Dong, Yinan Gao, Hang He, Jiajun Chai, Guojun Yin, Wei Lin
https://arxiv.org/abs/2508.10293 ht…

Promoting Efficient Reasoning with Verifiable Stepwise Reward
Large reasoning models (LRMs) have recently achieved significant progress in complex reasoning tasks, aided by reinforcement learning with verifiable rewards. However, LRMs often suffer from overthinking, expending excessive computation on simple problems and reducing efficiency. Existing efficient reasoning methods typically require accurate task assessment to preset token budgets or select reasoning modes, which limits their flexibility and reliability. In this work, we revisit the essence of…

OpenCodeReasoning-II: A Simple Test Time Scaling Approach via Self-Critique
Wasi Uddin Ahmad, Somshubra Majumdar, Aleksander Ficek, Sean Narenthiran, Mehrzad Samadi, Jocelyn Huang, Siddhartha Jain, Vahid Noroozi, Boris Ginsburg
https://arxiv.org/abs/2507.09075

OpenCodeReasoning-II: A Simple Test Time Scaling Approach via Self-Critique
Recent advancements in reasoning-based Large Language Models (LLMs), particularly their potential through test-time scaling, have created significant opportunities for distillation in code generation and critique. However, progress in both areas fundamentally depends on large-scale, high-quality datasets. In this work, we introduce OpenCodeReasoning-II, a dataset consists of 2.5M question-solution-critique triples (approx. 35K unique programming questions), making it nearly twice the size of th…

I mean, I don't usually write happy songs, but this looks good so far, right?

Judgment as Coordination: A Joint Systems View of Visualization Design Practice
Paul C. Parsons, Arran Ridley
https://arxiv.org/abs/2507.01209 https://

Judgment as Coordination: A Joint Systems View of Visualization Design Practice
Professional visualization design has become an increasingly important area of inquiry, yet much of the field's discourse remains anchored in researcher-centered contexts. Studies of design practice often focus on individual designers' decisions and reflections, offering limited insight into the collaborative and systemic dimensions of professional work. In this paper, we propose a systems-level reframing of design judgment grounded in the coordination and adaptation that sustain progress amid …

Hydrodynamic Effects in Cryogenic Buffer Gas Cells: Design Insights from Hybrid Simulations
Nick Vogeley, Bernd Bauerhenne, Daqing Wang
https://arxiv.org/abs/2508.04364 https://…

Hydrodynamic Effects in Cryogenic Buffer Gas Cells: Design Insights from Hybrid Simulations
Cryogenic buffer gas beam sources have become an essential tool for experiments requiring cold molecular beams with low forward velocities. Although recent experimental advances have led to significant progress in source optimization, numerical studies remain limited due to the challenges posed by the large parameter ranges required to describe both the dense buffer gas and the dilute seed molecules. In this work, we report a numerical evaluation of cryogenic buffer gas beam cells operating in …

During work (no technical hanky panky)

Firing up SWTOR on my Linux rig for tonight's stream! My character's Force powers are strong, but my ability to avoid standing in fire is... a work in progress. Wish me luck (and minimal rezzing)! #MMORPG #Gaming 🐧 🧙‍♂️

The Graph Structure of a Class of Permutation Maps over Ring $\mathbb{Z}_{p^k}$
Kai Tan, Chengqing Li
https://arxiv.org/abs/2506.20118 https://

The Graph Structure of a Class of Permutation Maps over Ring $\mathbb{Z}_{p^k}$
Understanding the periodic and structural properties of permutation maps over residue rings such as $\mathbb{Z}_{p^k}$ is a foundational challenge in algebraic dynamics and pseudorandom sequence analysis. Despite notable progress in characterizing global periods, a critical bottleneck remains: the lack of explicit tools to analyze local cycle structures and their evolution with increasing arithmetic precision. In this work, we propose a unified analytical framework to systematically derive the …

Exponential mixing for the stochastic Kuramoto-Sivashinsky equation on the 1D torus
Peng Gao, Hung D. Nguyen
https://arxiv.org/abs/2508.01794 https://arxiv…

Exponential mixing for the stochastic Kuramoto-Sivashinsky equation on the 1D torus
In this paper, we study the large-time behaviors of the Kuramoto-Sivashinsky equation (KSE) on the 1D torus while being subjected to random perturbation via additive Gaussian noise. It is well-known that under suitable assumptions on the stochastic forcing, the KSE admits a unique invariant probability measure. In this work, we make further progress on the topic of ergodicity by addressing the issue of convergence rate toward equilibrium. In comparison with the previous results, we can prove th…

Constraint Maps: Insights and Related Themes
Alessio Figalli, Andr\'e Guerra, Sunghan Kim, Henrik Shahgholian
https://arxiv.org/abs/2506.23608 https://…

Constraint Maps: Insights and Related Themes
This paper explores recent progress related to constraint maps. Building on the exposition in [14], our goal is to provide a clear and accessible account of some of the more intricate arguments behind the main results in this work. Along the way, we include several new results of independent value. In particular, we give optimal geometric conditions on the target manifold that guarantee a unique continuation result for the projected image map. We also prove that the gradient of a minimizing har…

Echo-4o: Harnessing the Power of GPT-4o Synthetic Images for Improved Image Generation
Junyan Ye, Dongzhi Jiang, Zihao Wang, Leqi Zhu, Zhenghao Hu, Zilong Huang, Jun He, Zhiyuan Yan, Jinghua Yu, Hongsheng Li, Conghui He, Weijia Li
https://arxiv.org/abs/2508.09987

Echo-4o: Harnessing the Power of GPT-4o Synthetic Images for Improved Image Generation
Recently, GPT-4o has garnered significant attention for its strong performance in image generation, yet open-source models still lag behind. Several studies have explored distilling image data from GPT-4o to enhance open-source models, achieving notable progress. However, a key question remains: given that real-world image datasets already constitute a natural source of high-quality data, why should we use GPT-4o-generated synthetic data? In this work, we identify two key advantages of syntheti…

WIP: Large Language Model-Enhanced Smart Tutor for Undergraduate Circuit Analysis
Liangliang Chen, Huiru Xie, Jacqueline Rohde, Ying Zhang
https://arxiv.org/abs/2506.08962

WIP: Large Language Model-Enhanced Smart Tutor for Undergraduate Circuit Analysis
This research-to-practice work-in-progress (WIP) paper presents an AI-enabled smart tutor designed to provide homework assessment and feedback for students in an undergraduate circuit analysis course. We detail the tutor's design philosophy and core components, including open-ended question answering and homework feedback generation. The prompts are carefully crafted to optimize responses across different problems. The smart tutor was deployed on the Microsoft Azure platform and is currently in…

Dyna-Think: Synergizing Reasoning, Acting, and World Model Simulation in AI Agents
Xiao Yu, Baolin Peng, Ruize Xu, Michel Galley, Hao Cheng, Suman Nath, Jianfeng Gao, Zhou Yu
https://arxiv.org/abs/2506.00320

Dyna-Think: Synergizing Reasoning, Acting, and World Model Simulation in AI Agents
Recent progress in reasoning with large language models (LLMs), such as DeepSeek-R1, demonstrates impressive capabilities in domains like mathematics and coding, by exhibiting complex cognitive behaviors such as verification, goal decomposition, and self-reflection. However, it is unclear what behavior is effective and what behavior is missing for long-horizon AI agents tasks. In this work, we propose Dyna-Think, a thinking framework that integrates planning with an internal world model with re…

Boosting Parameter Efficiency in LLM-Based Recommendation through Sophisticated Pruning
Shanle Zheng, Keqin Bao, Jizhi Zhang, Yang Zhang, Fuli Feng, Xiangnan He
https://arxiv.org/abs/2507.07064

Boosting Parameter Efficiency in LLM-Based Recommendation through Sophisticated Pruning
LLM-based recommender systems have made significant progress; however, the deployment cost associated with the large parameter volume of LLMs still hinders their real-world applications. This work explores parameter pruning to improve parameter efficiency while maintaining recommendation quality, thereby enabling easier deployment. Unlike existing approaches that focus primarily on inter-layer redundancy, we uncover intra-layer redundancy within components such as self-attention and MLP modules…

ChemDFM-R: An Chemical Reasoner LLM Enhanced with Atomized Chemical Knowledge
Zihan Zhao, Bo Chen, Ziping Wan, Lu Chen, Xuanze Lin, Shiyang Yu, Situo Zhang, Da Ma, Zichen Zhu, Danyang Zhang, Huayang Wang, Zhongyang Dai, Liyang Wen, Xin Chen, Kai Yu
https://arxiv.org/abs/2507.21990

ChemDFM-R: An Chemical Reasoner LLM Enhanced with Atomized Chemical Knowledge
While large language models (LLMs) have achieved impressive progress, their application in scientific domains such as chemistry remains hindered by shallow domain understanding and limited reasoning capabilities. In this work, we focus on the specific field of chemistry and develop a Chemical Reasoner LLM, ChemDFM-R. We first construct a comprehensive dataset of atomized knowledge points to enhance the model's understanding of the fundamental principles and logical structure of chemistry. Then,…

This https://arxiv.org/abs/2503.09532 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csLG_…

SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability
Sparse autoencoders (SAEs) are a popular technique for interpreting language model activations, and there is extensive recent work on improving SAE effectiveness. However, most prior work evaluates progress using unsupervised proxy metrics with unclear practical relevance. We introduce SAEBench, a comprehensive evaluation suite that measures SAE performance across eight diverse metrics, spanning interpretability, feature disentanglement and practical applications like unlearning. To enable syst…

Pride progress? As gay pro athletes consider coming out, each announcement makes a mark https://www.nytimes.com/athletic/6396179/2025/06/03/pride-month-gay-pro-athletes-evolution/

DreamAnywhere: Object-Centric Panoramic 3D Scene Generation
Edoardo Alberto Dominici, Jozef Hladky, Floor Verhoeven, Lukas Radl, Thomas Deixelberger, Stefan Ainetter, Philipp Drescher, Stefan Hauswiesner, Arno Coomans, Giacomo Nazzaro, Konstantinos Vardis, Markus Steinberger
https://arxiv.org/abs/2506.20367

DreamAnywhere: Object-Centric Panoramic 3D Scene Generation
Recent advances in text-to-3D scene generation have demonstrated significant potential to transform content creation across multiple industries. Although the research community has made impressive progress in addressing the challenges of this complex task, existing methods often generate environments that are only front-facing, lack visual fidelity, exhibit limited scene understanding, and are typically fine-tuned for either indoor or outdoor settings. In this work, we address these issues and …

You can use three.js as GIS to render contour lines from a Geotiff with customizable labels! Work in progress.
(P.S. can you tell where this might be? ;) ) #gischat

Post-AGB Binaries as Interacting Systems
Hans Van Winckel
https://arxiv.org/abs/2507.02514 https://arxiv.org/pdf/2507.02514

Post-AGB Binaries as Interacting Systems
We present recent progress in our understanding of the physical interaction mechanisms at work in evolved binaries of low-to-intermediate initial mass, which are surrounded by a stable disc of gas and dust. These systems are known as post-asymptotic giant-branch (post-AGB) binaries, but recently, it has been shown that some systems are too low in luminosity and should be considered as post-red-giant branch (post-RGB) instead. While the systems are currently well within their Roche lobe, they st…

Rethinking Backbone Design for Lightweight 3D Object Detection in LiDAR
Adwait Chandorkar, Hasan Tercan, Tobias Meisen
https://arxiv.org/abs/2508.00744 https://

Rethinking Backbone Design for Lightweight 3D Object Detection in LiDAR
Recent advancements in LiDAR-based 3D object detection have significantly accelerated progress toward the realization of fully autonomous driving in real-world environments. Despite achieving high detection performance, most of the approaches still rely on a VGG-based or ResNet-based backbone for feature exploration, which increases the model complexity. Lightweight backbone design is well-explored for 2D object detection, but research on 3D object detection still remains limited. In this work,…

First-principles phonon physics using the Pheasy code
Changpeng Lin, Jian Han, Ben Xu, Nicola Marzari
https://arxiv.org/abs/2508.01020 https://arxiv.org/pd…

First-principles phonon physics using the Pheasy code
Parameter-free calculations of lattice dynamics from first principles have achieved significant progress in the past decades, with a wealth of applications in thermodynamics, phase transitions, and transport properties of materials. Current approaches to derive the interatomic force constants (IFCs) of lattice potential become challenging and sometimes infeasible when going beyond third-order anharmonicity, due to the combinatorial explosion in the number of higher-order IFCs. In this work, we …

An Update on the Raiders' Training Camp Progress https://www.si.com/nfl/raiders/las-vegas-geno-smith-training-camp-jakobi-meyers-pete-carroll-brock-bowers

Repeating Flares, X-ray Outbursts and Delayed Infrared Emission: A Comprehensive Compilation of Optical Tidal Disruption Events
D. A. Langis, I. Liodakis, K. I. I. Koljonen, A. Paggi, N. Globus, L. Wyrzykowski, P. J. Miko{\l}ajczyk, K. Kotysz, P. Zieli\'nski, N. Ihanec, J. Ding, D. Morshed, Z. Torres
https://arxiv.org/abs/25…

Repeating Flares, X-ray Outbursts and Delayed Infrared Emission: A Comprehensive Compilation of Optical Tidal Disruption Events
Tidal disruption events (TDEs) have been proposed as valuable laboratories for studying dormant black holes. However, progress in this field has been hampered by the limited number of observed events. In this work, we present TDECat, a comprehensive catalogue of 134 confirmed TDEs (131 optical TDEs and 3 jetted TDEs) discovered up to the end of 2024, accompanied by multi-wavelength photometry (X-ray, UV, optical, and IR) and publicly available spectra. We also study the statistical properties, …

"In a world that demands spectacle, moss reminds us: softness can be a strategy." -- more @ #activism

the season of small things
how to trust the quiet work that doesn't look like progress yet

AccessGuru: Leveraging LLMs to Detect and Correct Web Accessibility Violations in HTML Code
Nadeen Fathallah, Daniel Hern\'andez, Steffen Staab
https://arxiv.org/abs/2507.19549

AccessGuru: Leveraging LLMs to Detect and Correct Web Accessibility Violations in HTML Code
The vast majority of Web pages fail to comply with established Web accessibility guidelines, excluding a range of users with diverse abilities from interacting with their content. Making Web pages accessible to all users requires dedicated expertise and additional manual efforts from Web page providers. To lower their efforts and promote inclusiveness, we aim to automatically detect and correct Web accessibility violations in HTML code. While previous work has made progress in detecting certain…

Multimodal Mathematical Reasoning with Diverse Solving Perspective
Wenhao Shi, Zhiqiang Hu, Yi Bin, Yang Yang, See-Kiong Ng, Heng Tao Shen
https://arxiv.org/abs/2507.02804

Multimodal Mathematical Reasoning with Diverse Solving Perspective
Recent progress in large-scale reinforcement learning (RL) has notably enhanced the reasoning capabilities of large language models (LLMs), especially in mathematical domains. However, current multimodal LLMs (MLLMs) for mathematical reasoning often rely on one-to-one image-text pairs and single-solution supervision, overlooking the diversity of valid reasoning perspectives and internal reflections. In this work, we introduce MathV-DP, a novel dataset that captures multiple diverse solution tra…

Time-Masked Transformers with Lightweight Test-Time Adaptation for Neural Speech Decoding
Ebrahim Feghhi, Shreyas Kaasyap, Nima Hadidi, Jonathan C. Kao
https://arxiv.org/abs/2507.02800

Time-Masked Transformers with Lightweight Test-Time Adaptation for Neural Speech Decoding
Speech neuroprostheses aim to restore communication for people with severe paralysis by decoding speech directly from neural activity. To accelerate algorithmic progress, a recent benchmark released intracranial recordings from a paralyzed participant attempting to speak, along with a baseline decoding algorithm. Prior work on the benchmark showed impressive accuracy gains. However, these gains increased computational costs and were not demonstrated in a real-time decoding setting. Here, we mak…

GEAR: Gaze-Enabled Human-Robot Collaborative Assembly
Asad Ali Shahid, Angelo Moroncelli, Drazen Brscic, Takayuki Kanda, Loris Roveda
https://arxiv.org/abs/2507.18947 https://…

GEAR: Gaze-Enabled Human-Robot Collaborative Assembly
Recent progress in robot autonomy and safety has significantly improved human-robot interactions, enabling robots to work alongside humans on various tasks. However, complex assembly tasks still present significant challenges due to inherent task variability and the need for precise operations. This work explores deploying robots in an assistive role for such tasks, where the robot assists by fetching parts while the skilled worker provides high-level guidance and performs the assembly. We intr…

Nonparametric Reaction Coordinate Optimization with Histories: A Framework for Rare Event Dynamics
Polina V. Banushkina, Sergei V. Krivov
https://arxiv.org/abs/2508.07326 https:…

Nonparametric Reaction Coordinate Optimization with Histories: A Framework for Rare Event Dynamics
Rare but critical events in complex systems, such as protein folding, chemical reactions, disease progression, and extreme weather or climate phenomena, are governed by complex, high-dimensional, stochastic dynamics. Identifying an optimal reaction coordinate (RC) that accurately captures the progress of these dynamics is crucial for understanding and simulating such processes. This work introduces a nonparametric RC optimization framework that incorporates trajectory histories, enabling robust…

RAAG: Ratio Aware Adaptive Guidance
Shangwen Zhu, Qianyu Peng, Yuting Hu, Zhantao Yang, Han Zhang, Zhao Pu, Ruili Feng, Fan Cheng
https://arxiv.org/abs/2508.03442 https://

RAAG: Ratio Aware Adaptive Guidance
Flow-based generative models have recently achieved remarkable progress in image and video synthesis, with classifier-free guidance (CFG) becoming the standard tool for high-fidelity, controllable generation. However, despite their practical success, little is known about how guidance interacts with different stages of the sampling process-especially in the fast, low-step regimes typical of modern flow-based pipelines. In this work, we uncover and analyze a fundamental instability: the earliest…

WebArXiv: Evaluating Multimodal Agents on Time-Invariant arXiv Tasks
Zihao Sun, Meng Fang, Ling Chen
https://arxiv.org/abs/2507.00938 https://

WebArXiv: Evaluating Multimodal Agents on Time-Invariant arXiv Tasks
Recent progress in large language models (LLMs) has enabled the development of autonomous web agents capable of navigating and interacting with real websites. However, evaluating such agents remains challenging due to the instability and inconsistency of existing benchmarks, which often rely on dynamic content or oversimplified simulations. In this work, we introduce WebArXiv, a static and time-invariant benchmark comprising 275 web-based tasks grounded in the arXiv platform. WebArXiv ensures r…

Joint ASR and Speaker Role Tagging with Serialized Output Training
Anfeng Xu, Tiantian Feng, Shrikanth Narayanan
https://arxiv.org/abs/2506.10349 https://

Joint ASR and Speaker Role Tagging with Serialized Output Training
Automatic Speech Recognition systems have made significant progress with large-scale pre-trained models. However, most current systems focus solely on transcribing the speech without identifying speaker roles, a function that is critical for conversational AI. In this work, we investigate the use of serialized output training (SOT) for joint ASR and speaker role tagging. By augmenting Whisper with role-specific tokens and fine-tuning it with SOT, we enable the model to generate role-aware trans…

Dissipative Coupling in Photonic and Plasmonic Resonators
Tong Wu, Philippe Lalanne
https://arxiv.org/abs/2507.20132 https://arxiv.org/pdf/2507.20132

Dissipative Coupling in Photonic and Plasmonic Resonators
The rapid progress of nanophotonics demands theoretical frameworks capable of predicting the resonant behavior of complex systems comprising constituents of varying nature, operating beyond the weak-coupling, high-Q regime where classical temporal coupled-mode theory (CMT) is applicable. This work presents a coupled-quasinormal-mode (cQNM) framework for analyzing dissipative coupling with photonic and plasmonic resonators. The framework provides rigorous closed-form expressions for dissipative …

Prometheus: Unified Knowledge Graphs for Issue Resolution in Multilingual Codebases
Zimin Chen, Yue Pan, Siyu Lu, Jiayi Xu, Claire Le Goues, Martin Monperrus, He Ye
https://arxiv.org/abs/2507.19942

Prometheus: Unified Knowledge Graphs for Issue Resolution in Multilingual Codebases
Language model (LM) agents, such as SWE-agent and OpenHands, have made progress toward automated issue resolution. However, existing approaches are often limited to Python-only issues and rely on pre-constructed containers in SWE-bench with reproduced issues, restricting their applicability to real-world and work for multi-language repositories. We present Prometheus, designed to resolve real-world issues beyond benchmark settings. Prometheus is a multi-agent system that transforms an entire co…

URSA: The Universal Research and Scientific Agent
Michael Grosskopf, Russell Bent, Rahul Somasundaram, Isaac Michaud, Arthur Lui, Nathan Debardeleben, Earl Lawrence
https://arxiv.org/abs/2506.22653

URSA: The Universal Research and Scientific Agent
Large language models (LLMs) have moved far beyond their initial form as simple chatbots, now carrying out complex reasoning, planning, writing, coding, and research tasks. These skills overlap significantly with those that human scientists use day-to-day to solve complex problems that drive the cutting edge of research. Using LLMs in "agentic" AI has the potential to revolutionize modern science and remove bottlenecks to progress. In this work, we present URSA, a scientific agent ecosystem for…

Adversarial Defence without Adversarial Defence: Enhancing Language Model Robustness via Instance-level Principal Component Removal
Yang Wang, Chenghao Xiao, Yizhi Li, Stuart E. Middleton, Noura Al Moubayed, Chenghua Lin
https://arxiv.org/abs/2507.21750

Adversarial Defence without Adversarial Defence: Enhancing Language Model Robustness via Instance-level Principal Component Removal
Pre-trained language models (PLMs) have driven substantial progress in natural language processing but remain vulnerable to adversarial attacks, raising concerns about their robustness in real-world applications. Previous studies have sought to mitigate the impact of adversarial attacks by introducing adversarial perturbations into the training process, either implicitly or explicitly. While both strategies enhance robustness, they often incur high computational costs. In this work, we propose …

Shape-for-Motion: Precise and Consistent Video Editing with 3D Proxy
Yuhao Liu, Tengfei Wang, Fang Liu, Zhenwei Wang, Rynson W. H. Lau
https://arxiv.org/abs/2506.22432

Shape-for-Motion: Precise and Consistent Video Editing with 3D Proxy
Recent advances in deep generative modeling have unlocked unprecedented opportunities for video synthesis. In real-world applications, however, users often seek tools to faithfully realize their creative editing intentions with precise and consistent control. Despite the progress achieved by existing methods, ensuring fine-grained alignment with user intentions remains an open and challenging problem. In this work, we present Shape-for-Motion, a novel framework that incorporates a 3D proxy for …

Simulating Human Behavior with the Psychological-mechanism Agent: Integrating Feeling, Thought, and Action
Qing Dong, Pengyuan Liu, Dong Yu, Chen Kang
https://arxiv.org/abs/2507.19495

Simulating Human Behavior with the Psychological-mechanism Agent: Integrating Feeling, Thought, and Action
Generative agents have made significant progress in simulating human behavior, but existing frameworks often simplify emotional modeling and focus primarily on specific tasks, limiting the authenticity of the simulation. Our work proposes the Psychological-mechanism Agent (PSYA) framework, based on the Cognitive Triangle (Feeling-Thought-Action), designed to more accurately simulate human behavior. The PSYA consists of three core modules: the Feeling module (using a layer model of affect to sim…

FlowETL: An Autonomous Example-Driven Pipeline for Data Engineering
Mattia Di Profio, Mingjun Zhong, Yaji Sripada, Marcel Jaspars
https://arxiv.org/abs/2507.23118 https://

FlowETL: An Autonomous Example-Driven Pipeline for Data Engineering
The Extract, Transform, Load (ETL) workflow is fundamental for populating and maintaining data warehouses and other data stores accessed by analysts for downstream tasks. A major shortcoming of modern ETL solutions is the extensive need for a human-in-the-loop, required to design and implement context-specific, and often non-generalisable transformations. While related work in the field of ETL automation shows promising progress, there is a lack of solutions capable of automatically designing a…

A Systematic Review of Human-AI Co-Creativity
Saloni Singh, Koen Hndriks, Drik Heylen, Kim Baraka
https://arxiv.org/abs/2506.21333 https://

A Systematic Review of Human-AI Co-Creativity
The co creativity community is making significant progress in developing more sophisticated and tailored systems to support and enhance human creativity. Design considerations from prior work can serve as a valuable and efficient foundation for future systems. To support this effort, we conducted a systematic literature review of 62 papers on co-creative systems. These papers cover a diverse range of applications, including visual arts, design, and writing, where the AI acts not just as a tool …

Active Inference AI Systems for Scientific Discovery
Karthik Duraisamy
https://arxiv.org/abs/2506.21329 https://arxiv.org/pdf/2506.21…

Active Inference AI Systems for Scientific Discovery
The rapid evolution of artificial intelligence has led to expectations of transformative scientific discovery, yet current systems remain fundamentally limited by their operational architectures, brittle reasoning mechanisms, and their separation from experimental reality. Building on earlier work, we contend that progress in AI-driven science now depends on closing three fundamental gaps -- the abstraction gap, the reasoning gap, and the reality gap -- rather than on model size/data/test time …

Setting The Table with Intent: Intent-aware Schema Generation and Editing for Literature Review Tables
Vishakh Padmakumar, Joseph Chee Chang, Kyle Lo, Doug Downey, Aakanksha Naik
https://arxiv.org/abs/2507.19521

Setting The Table with Intent: Intent-aware Schema Generation and Editing for Literature Review Tables
The increasing volume of academic literature makes it essential for researchers to organize, compare, and contrast collections of documents. Large language models (LLMs) can support this process by generating schemas defining shared aspects along which to compare papers. However, progress on schema generation has been slow due to: (i) ambiguity in reference-based evaluations, and (ii) lack of editing/refinement methods. Our work is the first to address both issues. First, we present an approach…