Tootfinder

"World’s first Indigenous-led ocean reserve is one step closer to reality"
#Oceans #Environment

World’s first Indigenous-led ocean reserve is one step closer to reality | The Optimist Daily
A visionary project blending Indigenous knowledge and science could create the first marine reserve led by Pacific Island nations.

Realism Control One-step Diffusion for Real-World Image Super-Resolution
Zongliang Wu, Siming Zheng, Peng-Tao Jiang, Xin Yuan
https://arxiv.org/abs/2509.10122 https://

Realism Control One-step Diffusion for Real-World Image Super-Resolution
Pre-trained diffusion models have shown great potential in real-world image super-resolution (Real-ISR) tasks by enabling high-resolution reconstructions. While one-step diffusion (OSD) methods significantly improve efficiency compared to traditional multi-step approaches, they still have limitations in balancing fidelity and realism across diverse scenarios. Since the OSDs for SR are usually trained or distilled by a single timestep, they lack flexible control mechanisms to adaptively prioriti…

Joe Burrow turf toe: Who is Jake Browning, what does he bring, and who needs to step up for Bengals?

https://www.cbssports.com/nfl/news/joe-bur

Raiders’ Win Over Titans: Step Forward or Fool’s Gold? https://raiderramble.com/2025/10/14/raiders-win-over-titans-step-forward-or-fools-gold/

Sources: UK security services including MI5 step up work with the country's largest companies over hacking fears and to improve UK cyber warfare preparedness (Financial Times)
https://www.ft.com/content/6e54967b-48ce-4788-99c7-36b5801e7dfe

UK security services step up work with business to fight cyber threats
High-profile hacks raise fears about the potential economic damage caused by disruption to supply chains and services

ESCoT: An Enhanced Step-based Coordinate Trajectory Planning Method for Multiple Car-like Robots
Junkai Jiang, Yihe Chen, Yibin Yang, Ruochen Li, Shaobing Xu, Jianqiang Wang
https://arxiv.org/abs/2508.09581

ESCoT: An Enhanced Step-based Coordinate Trajectory Planning Method for Multiple Car-like Robots
Multi-vehicle trajectory planning (MVTP) is one of the key challenges in multi-robot systems (MRSs) and has broad applications across various fields. This paper presents ESCoT, an enhanced step-based coordinate trajectory planning method for multiple car-like robots. ESCoT incorporates two key strategies: collaborative planning for local robot groups and replanning for duplicate configurations. These strategies effectively enhance the performance of step-based MVTP methods. Through extensive ex…

🇺🇦 #NowPlaying on KEXP's #VarietyMix
ODF:
🎵 Deep Step
#ODF
https://shallnotfade.co.uk/album/deep-step-ep

Sparse Polyak: an adaptive step size rule for high-dimensional M-estimation
Tianqi Qiao, Marie Maros
https://arxiv.org/abs/2509.09802 https://arxiv.org/pdf…

Sparse Polyak: an adaptive step size rule for high-dimensional M-estimation
We propose and study Sparse Polyak, a variant of Polyak's adaptive step size, designed to solve high-dimensional statistical estimation problems where the problem dimension is allowed to grow much faster than the sample size. In such settings, the standard Polyak step size performs poorly, requiring an increasing number of iterations to achieve optimal statistical precision-even when, the problem remains well conditioned and/or the achievable precision itself does not degrade with problem size.…

A central limit theorem for unbalanced step-reinforced random walks
Zhishui Hu, Liang Dong
https://arxiv.org/abs/2510.10898 https://arxiv.org/pdf/2510.1089…

A central limit theorem for unbalanced step-reinforced random walks
In this paper, we study a class of unbalanced step-reinforced random walks that unifies the elephant random walk, the positively step-reinforced random walk, and the negatively step-reinforced random walk. By establishing a connection with bond percolation on random recursive trees, these processes can be represented as randomly weighted sums of independent and identically distributed random variables. We first derive normal and stable central limit theorems for such randomly weighted sums, and…

From <Answer> to <Think>: Multidimensional Supervision of Reasoning Process for LLM Optimization
Beining Wang, Weihang Su, Hongtao Tian, Tao Yang, Yujia Zhou, Ting Yao, Qingyao Ai, Yiqun Liu
https://arxiv.org/abs/2510.11457

From to : Multidimensional Supervision of Reasoning Process for LLM Optimization
Improving the multi-step reasoning ability of Large Language Models (LLMs) is a critical yet challenging task. The dominant paradigm, outcome-supervised reinforcement learning (RLVR), rewards only correct final answers, often propagating flawed reasoning and suffering from sparse reward signals. While process-level reward models (PRMs) provide denser, step-by-step feedback, they lack generalizability and interpretability, requiring task-specific segmentation of the reasoning process. To this en…

Recognizing a Palestinian state is a limited but welcome step that addresses an enduring blind spot: Palestinian rights cannot be conditioned on Israeli interests. https://www.theguardian.com/us-news/ng-interactive/20…

Sensitivity dependence of the Navier-Stokes turbulence of a two-dimensional Rayleigh-B\'{e}nard convection on time-step
Shijie Qin, Kun Xu, Shijun Liao
https://arxiv.org/abs/2510.11220

Sensitivity dependence of the Navier-Stokes turbulence of a two-dimensional Rayleigh-Bénard convection on time-step
A two-dimensional Rayleigh-Bénard convection governed by the Navier-Stokes (NS) equations is solved by traditional direct numerical simulation (DNS) using double precision and various time-steps. It is found that there are two kinds of final flow types, one is vortical flow, the other is zonal flow, and their statistics are completely different. Especially, the two flow types frequently alternate as the time-step decreases to a very small value, suggesting that the time-step corresponding to e…

Reveal-Bangla: A Dataset for Cross-Lingual Multi-Step Reasoning Evaluation
Khondoker Ittehadul Islam, Gabriele Sarti
https://arxiv.org/abs/2508.08933 https://

Reveal-Bangla: A Dataset for Cross-Lingual Multi-Step Reasoning Evaluation
Language models have demonstrated remarkable performance on complex multi-step reasoning tasks. However, their evaluation has been predominantly confined to high-resource languages such as English. In this paper, we introduce a manually translated Bangla multi-step reasoning dataset derived from the English Reveal dataset, featuring both binary and non-binary question types. We conduct a controlled evaluation of English-centric and Bangla-centric multilingual small language models on the origin…

ARMA Block: A CNN-Based Autoregressive and Moving Average Module for Long-Term Time Series Forecasting
Myung Jin Kim, YeongHyeon Park, Il Dong Yun
https://arxiv.org/abs/2509.10324

ARMA Block: A CNN-Based Autoregressive and Moving Average Module for Long-Term Time Series Forecasting
This paper proposes a simple yet effective convolutional module for long-term time series forecasting. The proposed block, inspired by the Auto-Regressive Integrated Moving Average (ARIMA) model, consists of two convolutional components: one for capturing the trend (autoregression) and the other for refining local variations (moving average). Unlike conventional ARIMA, which requires iterative multi-step forecasting, the block directly performs multi-step forecasting, making it easily extendabl…

I'm enjoying the levity that the people arriving from bluesky are bringing, but they are too good at posting and taking over the feed. The rest of you need to step it up.

Every criticism of Israel,
any sympathy for the Palestinians,
any pressure to end the war,
is seen today as a form of antisemitism
– even endorsement of Hamas.
Over 60% of Israelis, according to recent polls, believe “nobody in Gaza is innocent.”
Along with Israel’s sense of impunity, this popular notion explains how this war has turned into a genocide.
The core of Palestinian identity, most Israelis believe, is the physical destruction of their stat…

From a bunch of csv files to a neat #SpatialAnalytics dataset:
Step-by-step GeoLife #GPS track collection processing with #DuckDB,

GeoLife GPS track collection processing with DuckDB, QGIS & Trajectools
The last time I preprocessed the whole GeoLife dataset, I loaded it into PostGIS. Today, I want to share a new workflow that creates a (Geo)Parquet file and that is much faster. The dataset (GeoLif…

Through-hole soldering test boards got here.
I have nine of them so plenty of opportunities for experiments. I think for the first step I'll stencil paste on one and reflow, then hand solder the second, and cross section them both.

GitHub CEO Thomas Dohmke plans to step down to pursue entrepreneurial endeavors; memo: Microsoft is opting not to directly replace the position (Ina Fried/Axios)
https://www.axios.com/2025/08/11/github-ceo-dohmke-step-down

Exclusive: GitHub CEO to step down
The Microsoft-owned software development hub has grown in importance in the AI era.

An optimal two-step estimation approach for two-phase studies
Qingning Zhou, Kin Yau Wong
https://arxiv.org/abs/2510.11587 https://arxiv.org/pdf/2510.11587…

An optimal two-step estimation approach for two-phase studies
Two-phase sampling is commonly adopted for reducing cost and improving estimation efficiency. In many two-phase studies, the outcome and some cheap covariates are observed for a large sample in Phase I, and expensive covariates are obtained for a selected subset of the sample in Phase II. As a result, the analysis of the association between the outcome and covariates faces a missing data problem. Complete-case analysis, which relies solely on the Phase II sample, is generally inefficient. In th…

Ignore all the diversions, the facts are Trump gave Putin a deadline with the promise of sanctions if he didn't comply. Putin increased his attacks on Ukraine, attacked Poland with drones. Putin didn't comply.
Trump's next step is demanding things from Europe..
#usa #trump

"Women in Mexico step up to protect the island farms traditionally inherited by men"
#Mexico #Women

Women in Mexico step up to protect the island farms traditionally inherited by men
This story is a collaboration between The Associated Press and Mongabay. MEXICO CITY, Mexico — Jasmín Ordóñez looks out at the water as the wooden boat she is in crosses a narrow channel that connects a labyrinth of chinampas, island farms that were built by the Aztecs thousands of years ago. “Let’s close our eyes […]

Spotify signs a direct licensing deal with rights group Kobalt to provide "greater flexibility, efficiency, value, and protections to songwriters in the US" (Ethan Millman/The Hollywood Reporter)
https://www.hollywoodreporter.com/news/mus

Spotify Secures Direct Licensing Deal with Kobalt
This partnership is a step in the right direction, and we look forward to continuing to work with Spotify to increase the value of songwriter royalties," Kobalt CEO Laurent Hubert said.

Positive Hermitian curvature flow on 2-step nilpotent Lie groups
Ettore Lo Giudice
https://arxiv.org/abs/2510.08846 https://arxiv.org/pdf/2510.08846…

Positive Hermitian curvature flow on 2-step nilpotent Lie groups
We study the positive Hermitian curvature flow for left-invariant metrics on $2$-step nilpotent Lie groups with a left-invariant complex structure $J$. We describe the long-time behavior of the flow under the assumption that $J[\mathfrak{g}, \mathfrak{g}]$ is contained in the center of $\mathfrak{g}$. We show that under our assumption the flow $g_{t}$ exists for all positive $t$ and $(G,(1+t)^{-1}g_{t})$ converges, in the Cheeger-Gromov topology, to a $2$-step nilpotent Lie group with a non fla…

i got up at 6 a.m. to wait in line

Linking EV charging and renewables in location and time unleashes additional benefits for the energy transition. How can we make the most of existing data and tools like smart charging?
In my new #RAP Regulatory Assistance Project brief, I explore the EU's REDIII and how countries can take the next step.
#EV

'Liberating our land, step by step' — Ukraine reports successful counterattack in Dnipropterovsk Oblast, shows footage: https://benborges.xyz/2025/10/07/liberating-our-land-step-by.html

Step 1) Someone (read manager) uses AI to put together a 30 lines python script.
Step 2) He declares in the next meeting that developers must become much more productive right away.
Step 3) … please tell me what you think will happen?

Read this thread all the way through and rate your enthusiasm for it.
https://mastodon.social/@pid_eins/115025535193319606

🪶 3D-printed superconductor achieves record performance with soft matter approach
#cornell …

3D-printed superconductor achieves record performance with soft matter approach
Nearly a decade after they first demonstrated that soft materials could guide the formation of superconductors, Cornell researchers have achieved a one-step, 3D printing method that produces superconductors with record properties.

STEP: Stepwise Curriculum Learning for Context-Knowledge Fusion in Conversational Recommendation
Zhenye Yang, Jinpeng Chen, Huan Li, Xiongnan Jin, Xuanyang Li, Junwei Zhang, Hongbo Gao, Kaimin Wei, Senzhang Wang
https://arxiv.org/abs/2508.10669

STEP: Stepwise Curriculum Learning for Context-Knowledge Fusion in Conversational Recommendation
Conversational recommender systems (CRSs) aim to proactively capture user preferences through natural language dialogue and recommend high-quality items. To achieve this, CRS gathers user preferences via a dialog module and builds user profiles through a recommendation module to generate appropriate recommendations. However, existing CRS faces challenges in capturing the deep semantics of user preferences and dialogue context. In particular, the efficient integration of external knowledge graph…

Optimum 1-Step Majority-Logic Decoding of Binary Reed-Muller Codes
Hoang Ly, Emina Soljanin
https://arxiv.org/abs/2508.08736 https://arxiv.org/pdf/2508.087…

Optimum 1-Step Majority-Logic Decoding of Binary Reed-Muller Codes
The classical majority-logic decoder proposed by Reed for Reed-Muller codes RM(r, m) of order r and length 2^m, unfolds in r+1 sequential steps, decoding message symbols from highest to lowest degree. Several follow-up decoding algorithms reduced the number of steps, but for a limited set of parameters, or at the expense of reduced performance, or relying on the existence of some combinatorial structures. We show that any one-step majority-logic decoder-that is, a decoder performing all majorit…

Breaking the Sabatier Principle by Dynamic Adsorption-Desorption Decoupling in Electrocatalytic Hydrogen Evolution
Zi-Xuan Yang, Lei Li, Tao Huang, Hui Wan, X. S. Wang, Gui-Fang Huang, Wangyu Hu, Wei-Qing Huang
https://arxiv.org/abs/2510.10555

Breaking the Sabatier Principle by Dynamic Adsorption-Desorption Decoupling in Electrocatalytic Hydrogen Evolution
The Sabatier principle establishes a fundamental trade-off in heterogeneous electrocatalysis.In the hydrogen evolution reaction (HER), this trade-off is manifested by the coupling of Volmer step, which requires strong hydrogen adsorption, with the Heyrovsky/Tafel step, which favors facile desorption, thus giving rise to the classical volcano relationship and limiting activity even at $ΔG=0$. Here, we demonstrate a ferroelectric platform with dynamic tunability -- monolayer GeS$_2$ decorated wi…

Adaptive Conditional Gradient Descent
Abbas Khademi, Antonio Silveti-Falls
https://arxiv.org/abs/2510.11440 https://arxiv.org/pdf/2510.11440

Adaptive Conditional Gradient Descent
Selecting an effective step-size is a fundamental challenge in first-order optimization, especially for problems with non-Euclidean geometries. This paper presents a novel adaptive step-size strategy for optimization algorithms that rely on linear minimization oracles, as used in the Conditional Gradient or non-Euclidean Normalized Steepest Descent algorithms. Using a simple heuristic to estimate a local Lipschitz constant for the gradient, we can determine step-sizes that guarantee sufficient …

US ethnic cleansing and what to do about it
Reposting link to source article instead of screenshot of tweet that had no alt text:
Data on arrests shows that ICE was heavily engaged in racial profiling in LA, because their arrest numbers fell by ~66% after that were ordered to stop making arrests based just in factors like skin color, with place, or language spoken.
#ICE #USPol

A Comprehensive Evaluation of Multilingual Chain-of-Thought Reasoning: Performance, Consistency, and Faithfulness Across Languages
Raoyuan Zhao, Yihong Liu, Hinrich Sch\"utze, Michael A. Hedderich
https://arxiv.org/abs/2510.09555

A Comprehensive Evaluation of Multilingual Chain-of-Thought Reasoning: Performance, Consistency, and Faithfulness Across Languages
Large reasoning models (LRMs) increasingly rely on step-by-step Chain-of-Thought (CoT) reasoning to improve task performance, particularly in high-resource languages such as English. While recent work has examined final-answer accuracy in multilingual settings, the thinking traces themselves, i.e., the intermediate steps that lead to the final answer, remain underexplored. In this paper, we present the first comprehensive study of multilingual CoT reasoning, evaluating three key dimensions: per…

Exact Classicalization of N-Level Quantum Systems Interacting with a Bath: Theory and Applications
Daniel Mart\'inez-Gil, Pedro Bargue\~no, Salvador Miret-Art\'es
https://arxiv.org/abs/2509.10131

Exact Classicalization of N-Level Quantum Systems Interacting with a Bath: Theory and Applications
In this manuscript, starting from a five-step algorithmic procedure for exactly classicalizing the dynamics of N-level quantum systems, we incorporate a classical bath of harmonic oscillators to model environmental interactions. Using the geometry of complex projective spaces CP(N-1) and a Langevin formalism, we obtain N-1 Hamilton's equations which encode both the quantum system and the bath degrees of freedom, representing a generalization of the Caldeira-Legget model in a complex projective …

Generating Compilers for Qubit Mapping and Routing
Abtin Molavi, Amanda Xu, Ethan Cecchetti, Swamit Tannu, Aws Albarghouthi
https://arxiv.org/abs/2508.10781 https://

Generating Compilers for Qubit Mapping and Routing
Quantum computers promise to solve important problems faster than classical computers, potentially unlocking breakthroughs in materials science, chemistry, and beyond. Optimizing compilers are key to realizing this potential, as they minimize expensive resource usage and limit error rates. A critical compilation step is qubit mapping and routing (QMR), which finds mappings from circuit qubits to qubits on a target device and plans instruction execution while satisfying the device's connectivity…

@… @… @… hmm… I wouldn’t trade a build step for a JavaScript server runtime dependency, long…

Prepared for the Unknown: Adapting AIOps Capacity Forecasting Models to Data Changes
Lorena Poenaru-Olaru, Wouter van 't Hof, Adrian Stando, Arkadiusz P. Trawinski, Eileen Kapel, Jan S. Rellermeyer, Luis Cruz, Arie van Deursen
https://arxiv.org/abs/2510.10320

Prepared for the Unknown: Adapting AIOps Capacity Forecasting Models to Data Changes
Capacity management is critical for software organizations to allocate resources effectively and meet operational demands. An important step in capacity management is predicting future resource needs often relies on data-driven analytics and machine learning (ML) forecasting models, which require frequent retraining to stay relevant as data evolves. Continuously retraining the forecasting models can be expensive and difficult to scale, posing a challenge for engineering teams tasked with balanc…

This Raiders Defender Will Take the Next Step This Season https://www.si.com/nfl/raiders/las-vegas-chris-smith-maxx-crosby-devin-white-pete-carroll

Missing Data Multiple Imputation for Tabular Q-Learning in Online RL
Kyla Chasalow, Skyler Wu, Susan Murphy
https://arxiv.org/abs/2510.10709 https://arxiv.…

Missing Data Multiple Imputation for Tabular Q-Learning in Online RL
Missing data in online reinforcement learning (RL) poses challenges compared to missing data in standard tabular data or in offline policy learning. The need to impute and act at each time step means that imputation cannot be put off until enough data exist to produce stable imputation models. It also means future data collection and learning depend on previous imputations. This paper proposes fully online imputation ensembles. We find that maintaining multiple imputation pathways may help bala…

Bills, the NFL's Anti-Cowboys, Continue to Sign Their Top Young Talent Early https://www.foxsports.com/stories/nfl/bills-nfls-anti-cowboys-continue-sign-top-young-talent-early

A GPU-resident Memory-Aware Algorithm for Accelerating Bidiagonalization of Banded Matrices
Evelyne Ringoot, Rabab Alomairy, Alan Edelman
https://arxiv.org/abs/2510.12705 https:…

A GPU-resident Memory-Aware Algorithm for Accelerating Bidiagonalization of Banded Matrices
The reduction of a banded matrix to a bidiagonal form is a crucial step in the Singular Value Decomposition (SVD), a cornerstone of scientific computing and AI. Despite being a highly parallel algorithm, it was previously believed to be unsuitable for GPU computation because it is memory bandwidth-bound. Recent developments in GPU hardware, including larger L1 memory per Streaming Multiprocessor/Compute Unit, have changed that. We present the first GPU algorithm for reducing a banded matrix to …

A Highly Predictive Neutrino Model: The Step Toward Precision
Pralay Chakraborty, Sagar Tirtha Goswami, Subhankar Roy
https://arxiv.org/abs/2508.07837 https://

A Highly Predictive Neutrino Model: The Step Toward Precision
A highly predictive neutrino mass matrix texture is proposed. The texture is minimal and carries two interesting correlations. It gives precise and unambiguous answers to almost all the current issues related to observable parameters in neutrino phenomenology. The texture is developed in a framework of type-I + type-II seesaw mechanisms with an extended group structure of $SU(2)_L \times U(1)_Y \times A_4 \times Z_{10} \times Z_{7} \times Z_{3}$.

For all the marketing advice LinkedIn MBAs come up with, a surprising number of people forget one of the most basic steps: The Middle Schooler Test.
The process is simple:
1) Place a sample of your brochure, website, etc. in front of a recently calibrated 12-year-old boy.
2) If they start giggling uncontrollably, identify and correct the source of their amusement. Return to step 1.
3) If they aren't particularly amused, move on with the campaign.

DEF CON volunteers step up to help water sector after China, Iran attack utilities https://therecord.media/def-con-franklin-water-utility-cybersecurity-volunteers

Transcendence Meets Normality: Construction of Transcendentally Normal Numbers
Chokri Manai
https://arxiv.org/abs/2508.09319 https://arxiv.org/pdf/2508.093…

Transcendence Meets Normality: Construction of Transcendentally Normal Numbers
In this work, we study real numbers $x$ for which $p(x)$ is (absolutely) normal for every non-constant integer-valued polynomial $p$. We call such numbers transcendentally normal. We prove that almost every real number is transcendentally normal and provide an explicit construction of such a number, based on Sierpinski's covering method and novel ideas involving the so-called stretch function. In the next step, we transform this construction into an algorithm that computes the digits of a t-nor…

DyKnow-RAG: Dynamic Knowledge Utilization Reinforcement Framework for Noisy Retrieval-Augmented Generation in E-commerce Search Relevance
Tingqiao Xu, Shaowei Yao, Chenhe Dong, Yiming Jin, Zerui Huang, Dan Ou, Haihong Tang
https://arxiv.org/abs/2510.11122

DyKnow-RAG: Dynamic Knowledge Utilization Reinforcement Framework for Noisy Retrieval-Augmented Generation in E-commerce Search Relevance
Accurately modeling query-item relevance drives e-commerce ranking, yet long-tail, knowledge-heavy, and fast-evolving queries exceed parametric LLM coverage. External context (reviews, attribute encyclopedias, UGC) can help but is noisy, and single-pass latency and cost forbid any clean-then-summarize step. The model must, per query, judge relevance and decide whether to use, partially use, or ignore the context. DyKnow-RAG is a dynamic noisy-RAG framework built on Group Relative Policy Optimiz…

Abduct, Act, Predict: Scaffolding Causal Inference for Automated Failure Attribution in Multi-Agent Systems
Alva West, Yixuan Weng, Minjun Zhu, Zhen Lin, Yue Zhang
https://arxiv.org/abs/2509.10401

Abduct, Act, Predict: Scaffolding Causal Inference for Automated Failure Attribution in Multi-Agent Systems
Failure attribution in multi-agent systems -- pinpointing the exact step where a decisive error occurs -- is a critical yet unsolved challenge. Current methods treat this as a pattern recognition task over long conversation logs, leading to critically low step-level accuracy (below 17\%), which renders them impractical for debugging complex systems. Their core weakness is a fundamental inability to perform robust counterfactual reasoning: to determine if correcting a single action would have ac…

Long time strong convergence analysis of one-step methods for McKean-Vlasov SDEs with superlinear growth coefficients
Taiyuan Liu, Yaozhong Hu, Siqing Gan
https://arxiv.org/abs/2509.09274

Long time strong convergence analysis of one-step methods for McKean-Vlasov SDEs with superlinear growth coefficients
This paper presents a strong convergence rate analysis of general discretization approximations for McKean-Vlasov SDEs with super-linear growth coefficients over infinite time horizon. Under some specified non-globally Lipschitz conditions, we derive the propagation of chaos, and the mean-square convergence rate over infinite time horizon for general one-step time discretization schemes for the underlying Mckean-Vlasov SDEs. As an application of the general result it is obtained the mean-square…

"At his party conference in Birmingham, 5 Houses* said Rayner’s actions “scream of entitlement”.
Crisis engulfs Labour as Angela Rayner is forced to step down as deputy PM | Labour | The Guardian
https://www.theguardian.com/politics/2025/…

A General Nonlinear Model for Arbitrary Modulation Formats in the Presence of Inter-Channel Simulated Raman Scattering
Zhiwei Liang, Bin Chen, Jiwei Xu, Yi Lei, Qingqing Hu, Fan Zhang, Gabriele Liga
https://arxiv.org/abs/2509.10009

A General Nonlinear Model for Arbitrary Modulation Formats in the Presence of Inter-Channel Simulated Raman Scattering
The four-dimensional nonlinear model is extended to include the inter-channel stimulated Raman scattering, enabling accurate prediction of dual-polarization four-dimensional modulation formats and probabilistically shaped constellations in high-dispersion regimes. The proposed model is validated via comparisons with the split-step Fourier method and enhanced Gaussian noise model.

I try not to step on anyone's toes, but peeps keep putting their feet under mine.

Cluster Topology-Driven Placement of Experts Reduces Network Traffic in MoE Inference
Danil Sivtsov, Aleksandr Katrutsa, Ivan Oseledets
https://arxiv.org/abs/2508.09229 https://…

Cluster Topology-Driven Placement of Experts Reduces Network Traffic in MoE Inference
Efficient deployment of a pre-trained LLM to a cluster with multiple servers is a critical step for providing fast responses to users' queries. The recent success of Mixture-of-Experts (MoE) LLMs raises the question of how to deploy them efficiently, considering their underlying structure. During the inference in MoE LLMs, only a small part of the experts is selected to process a given token. Moreover, in practice, the experts' load is highly imbalanced. For efficient deployment, one has to dis…

Structured identification of multivariable modal systems
Maarten van der Hulst, Rodrigo A. Gonz\'alez, Koen Classens, Paul Tacx, Nick Dirkx, Jeroen van de Wijdeven, Tom Oomen
https://arxiv.org/abs/2510.10820

Structured identification of multivariable modal systems
Physically interpretable models are essential for next-generation industrial systems, as these representations enable effective control, support design validation, and provide a foundation for monitoring strategies. The aim of this paper is to develop a system identification framework for estimating modal models of complex multivariable mechanical systems from frequency response data. To achieve this, a two-step structured identification algorithm is presented, where an additive model is first …

#CareerMoves
https://…

Our CTO, Ronald Snijder, has decided to take the next step in his career and will therefore be leaving the OAPEN Foundation at the end of this month. | Niels Stern | 16 comments
Our CTO, Ronald Snijder, has decided to take the next step in his career and will therefore be leaving the OAPEN Foundation at the end of this month. I have published a short farewell blog post here: https://lnkd.in/d4UW_k5p Wishing you all the best in your new role, Ronald! | 16 comments on LinkedIn

Holonomy of the Obata connection on 2-step hypercomplex nilmanifolds
Adri\'an Andrada, Mar\'ia Laura Barberis, Beatrice Brienza
https://arxiv.org/abs/2508.07889 https://…

Holonomy of the Obata connection on 2-step hypercomplex nilmanifolds
We study the holonomy of the Obata connection on 2-step hypercomplex nilmanifolds. By explicitly computing the curvature tensor, we determine the conditions under which the Obata connection is flat, showing that this depends on the nilpotency step of each complex structure. In particular, we show that the holonomy algebra of the Obata connection is always an abelian subalgebra of $\mathfrak{sl}(n, \mathbb{H})$: to establish this, we prove the $\mathbb{H}$-solvable conjecture for 2-step hypercom…

Phase Transitions for Elephant Random Walks with Two memory Channels
Krishanu Maulik, Parthanil Roy, Tamojit Sadhukhan
https://arxiv.org/abs/2509.10225 https://

Phase Transitions for Elephant Random Walks with Two memory Channels
Elephant random walk, introduced to study the effect of memory on random walks, is a novel type of walk that incorporates the information of one randomly chosen past step to determine the future step. However, memory of a process can be multifaceted and can arise due to interactions of more than one underlying phenomena. To model this, random walks with multiple memory channels were introduced in the statistical physics literature by Saha (2022) - here the information on a bunch of independentl…

Foraging with the Eyes: Dynamics in Human Visual Gaze and Deep Predictive Modeling
Tejaswi V. Panchagnula
https://arxiv.org/abs/2510.09299 https://arxiv.or…

Foraging with the Eyes: Dynamics in Human Visual Gaze and Deep Predictive Modeling
Animals often forage via Levy walks stochastic trajectories with heavy tailed step lengths optimized for sparse resource environments. We show that human visual gaze follows similar dynamics when scanning images. While traditional models emphasize image based saliency, the underlying spatiotemporal statistics of eye movements remain underexplored. Understanding these dynamics has broad applications in attention modeling and vision-based interfaces. In this study, we conducted a large scale huma…

NeST-BO: Fast Local Bayesian Optimization via Newton-Step Targeting of Gradient and Hessian Information
Wei-Ting Tang, Akshay Kudva, Joel A. Paulson
https://arxiv.org/abs/2510.05516

NeST-BO: Fast Local Bayesian Optimization via Newton-Step Targeting of Gradient and Hessian Information
Bayesian optimization (BO) is effective for expensive black-box problems but remains challenging in high dimensions. We propose NeST-BO, a local BO method that targets the Newton step by jointly learning gradient and Hessian information with Gaussian process surrogates, and selecting evaluations via a one-step lookahead bound on Newton-step error. We show that this bound (and hence the step error) contracts with batch size, so NeST-BO directly inherits inexact-Newton convergence: global progres…

Ready to share your story about how the Trump administration harmed you?
Just click the link.
https://ofthebrave.org/tell-your-story/

Tell Your Story - Home of the Brave
The first step in fighting back is speaking out.

Quantum kernel and HHL-based support vector machines for multi-class classification
Gabriela Pinheiro, Donovan Slabbert, Luis Kowada, Francesco Petruccione
https://arxiv.org/abs/2509.10190

Quantum kernel and HHL-based support vector machines for multi-class classification
We compare two quantum approaches that use support vector machines for multi-class classification on a reduced Sloan Digital Sky Survey (SDSS) dataset: the quantum kernel-based QSVM and the Harrow-Hassidim-Lloyd least-squares SVM (HHL LS-SVM). Both one-vs-rest and two-step hierarchical classification schemes were implemented. The QSVM involves angle encoding of ten features, two unitary operator blocks consisting of rotational operator gates, and a projective measurement that projects the final…

Hybrid Data-Driven Predictive Control for Robust and Reactive Exoskeleton Locomotion Synthesis
Kejun Li, Jeeseop Kim, Maxime Brunet, Marine P\'etriaux, Yisong Yue, Aaron D. Ames
https://arxiv.org/abs/2508.10269

Hybrid Data-Driven Predictive Control for Robust and Reactive Exoskeleton Locomotion Synthesis
Robust bipedal locomotion in exoskeletons requires the ability to dynamically react to changes in the environment in real time. This paper introduces the hybrid data-driven predictive control (HDDPC) framework, an extension of the data-enabled predictive control, that addresses these challenges by simultaneously planning foot contact schedules and continuous domain trajectories. The proposed framework utilizes a Hankel matrix-based representation to model system dynamics, incorporating step-to-…

Kahan's Automatic Step-Size Control for Unconstrained Optimization
Yifeng Meng, Chungen Shen, Linuo Xue, Lei-Hong Zhang
https://arxiv.org/abs/2508.06002 https://

Kahan's Automatic Step-Size Control for Unconstrained Optimization
The Barzilai and Borwein (BB) gradient method is one of the most widely-used line-search gradient methods. It computes the step-size for the current iterate by using the information carried in the previous iteration. Recently, William Kahan [Kahan, Automatic Step-Size Control for Minimization Iterations, Technical report, University of California, Berkeley CA, USA, 2019] proposed new Gradient Descent (KGD) step-size strategies which iterate the step-size itself by effectively utilizing the info…

SP-MoE: Speculative Decoding and Prefetching for Accelerating MoE-based Model Inference
Liangkun Chen, Zijian Wen, Tian Wu, Xiaoxi Zhang, Chuan Wu
https://arxiv.org/abs/2510.10302

SP-MoE: Speculative Decoding and Prefetching for Accelerating MoE-based Model Inference
The Mixture-of-Experts (MoE) architecture has been widely adopted in large language models (LLMs) to reduce computation cost through model sparsity. Employing speculative decoding (SD) can further accelerate MoE inference by drafting multiple tokens per step and verifying them in parallel. However, combining MoE with SD inflates GPU memory and aggravates CPU-GPU bandwidth contention during multi-token verification. Existing MoE offloading systems are SD-agnostic and do not address this bottlene…

Trump takes the next logical step toward dictatorship by turning the National Guard on US citizens in Washington.
Russia breaks through in the Donbas region, and things are not looking good for Ukraine.
Israel ramps up it’s genocide of Palestinians, and the death toll increases in Gaza. World leaders continue to turn a blind eye, offering only symbolic protest.
Perhaps the worst hunger crisis of all time in Sudan, and the world is not watching.
I have had quite enough…

Complex structures on two-step nilpotent Lie groups
Maria Laura Barberis
https://arxiv.org/abs/2508.05885 https://arxiv.org/pdf/2508.05885

Complex structures on two-step nilpotent Lie groups
We give a characterization of the $2$-step nilpotent Lie algebras whose corresponding Lie groups admit a left invariant complex structure. This is done by considering separately the cases when the complex structure is 2-step or 3-step nilpotent in the sense of [Cordero-Fernández-Gray-Ugarte, 2000]. We also study the Hermitian geometry 2-step nilpotent Lie groups. We show that if a left invariant Hermitian metric on such Lie group is pluriclosed, then the corresponding complex structure is 2-st…

"Step forward the ‘fully biodegradable’ shoe that leaves no footprint"
#Shoes #Fashion
…

Step forward the ‘fully biodegradable’ shoe that leaves no footprint
Made from organic materials and designed to return harmlessly to the earth, could this be the end of landfill-bound footwear?

Uncertainty Quantification for Retrieval-Augmented Reasoning
Heydar Soudani, Hamed Zamani, Faegheh Hasibi
https://arxiv.org/abs/2510.11483 https://arxiv.or…

Uncertainty Quantification for Retrieval-Augmented Reasoning
Retrieval-augmented reasoning (RAR) is a recent evolution of retrieval-augmented generation (RAG) that employs multiple reasoning steps for retrieval and generation. While effective for some complex queries, RAR remains vulnerable to errors and misleading outputs. Uncertainty quantification (UQ) offers methods to estimate the confidence of systems' outputs. These methods, however, often handle simple queries with no retrieval or single-step retrieval, without properly handling RAR setup. Accura…

Identifying & Interactively Refining Ambiguous User Goals for Data Visualization Code Generation
Mert \.Inan, Anthony Sicilia, Alex Xie, Saujas Vaduguru, Daniel Fried, Malihe Alikhani
https://arxiv.org/abs/2510.09390

Identifying & Interactively Refining Ambiguous User Goals for Data Visualization Code Generation
Establishing shared goals is a fundamental step in human-AI communication. However, ambiguities can lead to outputs that seem correct but fail to reflect the speaker's intent. In this paper, we explore this issue with a focus on the data visualization domain, where ambiguities in natural language impact the generation of code that visualizes data. The availability of multiple views on the contextual (e.g., the intended plot and the code rendering the plot) allows for a unique and comprehensive …

OMGSR: You Only Need One Mid-timestep Guidance for Real-World Image Super-Resolution
Zhiqiang Wu, Zhaomang Sun, Tong Zhou, Bingtao Fu, Ji Cong, Yitong Dong, Huaqi Zhang, Xuan Tang, Mingsong Chen, Xian Wei
https://arxiv.org/abs/2508.08227

OMGSR: You Only Need One Mid-timestep Guidance for Real-World Image Super-Resolution
Denoising Diffusion Probabilistic Models (DDPM) and Flow Matching (FM) generative models show promising potential for one-step Real-World Image Super-Resolution (Real-ISR). Recent one-step Real-ISR models typically inject a Low-Quality (LQ) image latent distribution at the initial timestep. However, a fundamental gap exists between the LQ image latent distribution and the Gaussian noisy latent distribution, limiting the effective utilization of generative priors. We observe that the noisy laten…

Simultaneous testing of hypotheses and alternatives
P. A. Koldanov, A. P. Koldanov
https://arxiv.org/abs/2509.10197 https://arxiv.org/pdf/2509.10197…

Simultaneous testing of hypotheses and alternatives
To identify statistically significant conclusions, it is proposed to simultaneously test hypotheses and alternatives. It is shown that, under the condition of free combination of hypotheses and alternatives, the closure method leads to single-step procedures for the simultaneous testing of hypotheses and alternatives. Using an example it is shown the result is lost when hypotheses and alternatives do not satisfy the condition of free combination of hypotheses and alternatives. It is shown that …

Boundary-Guided Policy Optimization for Memory-efficient RL of Diffusion Large Language Models
Nianyi Lin, Jiajie Zhang, Lei Hou, Juanzi Li
https://arxiv.org/abs/2510.11683 http…

Boundary-Guided Policy Optimization for Memory-efficient RL of Diffusion Large Language Models
A key challenge in applying reinforcement learning (RL) to diffusion large language models (dLLMs) lies in the intractability of their likelihood functions, which are essential for the RL objective, necessitating corresponding approximation in each training step. While existing methods approximate the log-likelihoods by their evidence lower bounds (ELBOs) via customized Monte Carlo (MC) sampling, the forward computational graphs of all MC samples need to be retained for the gradient computation…

Staggered time discretization in finitely-strained heterogeneous visco-elastodynamics with damage or diffusion in the Eulerian frame
Tom\'a\v{s} Roub\'i\v{c}ek
https://arxiv.org/abs/2510.10355 …

Staggered time discretization in finitely-strained heterogeneous visco-elastodynamics with damage or diffusion in the Eulerian frame
The semi-implicit (partly decoupled, also called staggered or fraction-step) time discretization is applied to compressible nonlinear dynamical models of viscoelastic solids in the Eulerian description, i.e.\ in the actual deforming configuration, formulated fully in terms of rates. The Kelvin-Voigt rheology and also, in the deviatoric part, the Jeffreys rheology are considered. The numerical stability and, considering the Stokes-type viscosity multipolar of the 2nd-grade, also convergence towa…

Convective flux analysis on the instability of one-dimensional detonation
Yunfeng Liu
https://arxiv.org/abs/2509.10169 https://arxiv.org/pdf/2509.10169

Convective flux analysis on the instability of one-dimensional detonation
One-dimensional numerical simulations using the Euler equations and irreversible one-step Arrhenius kinetics are conducted to study the instability mechanism of a one-dimensional gaseous detonation. By increasing the activation energy, this study identifies the characteristics of stable detonation, periodic detonation, pulsating detonation, and detonation quenching. The key difference between this study and previous research is that it is the first quantitative analysis of convective flux, kine…

Spectral theory for L\'evy and L\'evy-Ornstein-Uhlenbeck semigroups on step 2 Carnot groups
Maria Gordina, Rohan Sarkar
https://arxiv.org/abs/2510.08866 https://

Spectral theory for Lévy and Lévy-Ornstein-Uhlenbeck semigroups on step 2 Carnot groups
We consider non-local perturbations $Δ^ψ_G$ of sub-Laplacians on a step $2$ Carnot group $G$. The perturbations are by translation-invariant non-local operators acting along the vertical directions in $G$. We use harmonic analysis on $G$ to obtain intertwining relationship between the semigroups generated by $Δ^ψ_G$ and some strongly continuous contraction semigroups on Euclidean spaces with purely continuous spectrum, and as a result we identify the spectrum of $Δ^ψ_G$. Further we introd…

What Stood Out the Most About Raiders' Win Over Titans https://www.si.com/nfl/raiders/las-vegas-jeremy-chinn-tennessee-titans-pete-carroll

Scalable Training for Vector-Quantized Networks with 100% Codebook Utilization
Yifan Chang, Jie Qin, Limeng Qiao, Xiaofeng Wang, Zheng Zhu, Lin Ma, Xingang Wang
https://arxiv.org/abs/2509.10140

Scalable Training for Vector-Quantized Networks with 100% Codebook Utilization
Vector quantization (VQ) is a key component in discrete tokenizers for image generation, but its training is often unstable due to straight-through estimation bias, one-step-behind updates, and sparse codebook gradients, which lead to suboptimal reconstruction performance and low codebook usage. In this work, we analyze these fundamental challenges and provide a simple yet effective solution. To maintain high codebook usage in VQ networks (VQN) during learning annealing and codebook size expans…

Replaced article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[1/6]:
- Multi-Step Reasoning with Large Language Models, a Survey
Aske Plaat, Annie Wong, Suzan Verberne, Joost Broekens, Niki van Stein, Thomas Back

Offline Reinforcement Learning with Generative Trajectory Policies
Xinsong Feng, Leshu Tang, Chenan Wang, Haipeng Chen
https://arxiv.org/abs/2510.11499 https://

Offline Reinforcement Learning with Generative Trajectory Policies
Generative models have emerged as a powerful class of policies for offline reinforcement learning (RL) due to their ability to capture complex, multi-modal behaviors. However, existing methods face a stark trade-off: slow, iterative models like diffusion policies are computationally expensive, while fast, single-step models like consistency policies often suffer from degraded performance. In this paper, we demonstrate that it is possible to bridge this gap. The key to moving beyond the limitati…

ParallelSearch: Train your LLMs to Decompose Query and Search Sub-queries in Parallel with Reinforcement Learning
Shu Zhao, Tan Yu, Anbang Xu, Japinder Singh, Aaditya Shukla, Rama Akkiraju
https://arxiv.org/abs/2508.09303

ParallelSearch: Train your LLMs to Decompose Query and Search Sub-queries in Parallel with Reinforcement Learning
Reasoning-augmented search agents such as Search-R1, trained via reinforcement learning with verifiable rewards (RLVR), demonstrate remarkable capabilities in multi-step information retrieval from external knowledge sources. These agents address the limitations of their parametric memory by dynamically gathering relevant facts to address complex reasoning tasks. However, existing approaches suffer from a fundamental architectural limitation: they process search queries strictly sequentially, ev…

Perfectly transparent boundary conditions and wave propagation in lattice Boltzmann schemes
Thomas Bellotti (EM2C)
https://arxiv.org/abs/2509.10066 https://

Perfectly transparent boundary conditions and wave propagation in lattice Boltzmann schemes
Systems of N = 1, 2, . . . first-order hyperbolic conservation laws feature N undamped waves propagating at finite speeds. On their own hand, multi-step Finite Difference and lattice Boltzmann schemes with q = N + 1, N + 2, . . . unknowns involve N ''physical'' waves, which are aimed at being as closely-looking as possible to the ones of the PDEs, and q-N ''numerical-spurious-parasitic'' waves, which are subject to their own speed of propagation, and either damped or undamped. The whole picture…

Per-Query Visual Concept Learning
Ori Malca, Dvir Samuel, Gal Chechik
https://arxiv.org/abs/2508.09045 https://arxiv.org/pdf/2508.09045

Per-Query Visual Concept Learning
Visual concept learning, also known as Text-to-image personalization, is the process of teaching new concepts to a pretrained model. This has numerous applications from product placement to entertainment and personalized design. Here we show that many existing methods can be substantially augmented by adding a personalization step that is (1) specific to the prompt and noise seed, and (2) using two loss terms based on the self- and cross- attention, capturing the identity of the personalized co…

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[5/9]:
- Reveal-Bangla: A Dataset for Cross-Lingual Multi-Step Reasoning Evaluation
Khondoker Ittehadul Islam, Gabriele Sarti

Deep Learning to Identify the Spatio-Temporal Cascading Effects of Train Delays in a High-Density Network
Vu Duc Anh Nguyen, Ziyue Li
https://arxiv.org/abs/2510.09350 https://…

Deep Learning to Identify the Spatio-Temporal Cascading Effects of Train Delays in a High-Density Network
The operational efficiency of railway networks, a cornerstone of modern economies, is persistently undermined by the cascading effects of train delays. Accurately forecasting this delay propagation is a critical challenge for real-time traffic management. While recent research has leveraged Graph Neural Networks (GNNs) to model the network structure of railways, a significant gap remains in developing frameworks that provide multi-step autoregressive forecasts at a network-wide scale, while sim…

Towards Fully Automated Molecular Simulations: Multi-Agent Framework for Simulation Setup and Force Field Extraction
Marko Petkovi\'c, Vlado Menkovski, Sof\'ia Calero
https://arxiv.org/abs/2509.10210

Towards Fully Automated Molecular Simulations: Multi-Agent Framework for Simulation Setup and Force Field Extraction
Automated characterization of porous materials has the potential to accelerate materials discovery, but it remains limited by the complexity of simulation setup and force field selection. We propose a multi-agent framework in which LLM-based agents can autonomously understand a characterization task, plan appropriate simulations, assemble relevant force fields, execute them and interpret their results to guide subsequent steps. As a first step toward this vision, we present a multi-agent system…

Attention's forward pass and Frank-Wolfe
Albert Alcalde, Borjan Geshkovski, Dom\`enec Ruiz-Balet
https://arxiv.org/abs/2508.09628 https://arxiv.org/pdf…

Attention's forward pass and Frank-Wolfe
We study the hardmax limit of self-attention dynamics for token embeddings obtained in the zero-temperature ($β\to+\infty$) regime, and relate it to the finite-$β$ setting. In this limit, the update rule can be viewed as a Frank-Wolfe step for a quadratic objective over the convex hull of the current token embeddings. When the key-query matrix is negative semidefinite, the method linearly contracts all tokens to a single cluster at the origin. When it is positive semidefinite, extending the h…

How Raiders Can Take the Next Step Against the Chargers https://www.si.com/nfl/raiders/las-vegas-pete-carroll-los-angeles-chargers-chipkelly-geno-smith

RadioFlow: Efficient Radio Map Construction Framework with Flow Matching
Haozhe Jia, Wenshuo Chen, Xiucheng Wang, Nan Cheng, Hongbo Zhang, Kuimou Yu, Songning Lai, Nanjian Jia, Bowen Tian, Hongru Xiao, Yutao Yue
https://arxiv.org/abs/2510.09314

RadioFlow: Efficient Radio Map Construction Framework with Flow Matching
Accurate and real-time radio map (RM) generation is crucial for next-generation wireless systems, yet diffusion-based approaches often suffer from large model sizes, slow iterative denoising, and high inference latency, which hinder practical deployment. To overcome these limitations, we propose \textbf{RadioFlow}, a novel flow-matching-based generative framework that achieves high-fidelity RM generation through single-step efficient sampling. Unlike conventional diffusion models, RadioFlow lea…

Retrospective Sparse Attention for Efficient Long-Context Generation
Seonghwan Choi, Beomseok Kang, Dongwon Jo, Jae-Joon Kim
https://arxiv.org/abs/2508.09001 https://

Retrospective Sparse Attention for Efficient Long-Context Generation
Large Language Models (LLMs) are increasingly deployed in long-context tasks such as reasoning, code generation, and multi-turn dialogue. However, inference over extended contexts is bottlenecked by the Key-Value (KV) cache, whose memory footprint grows linearly with sequence length and dominates latency at each decoding step. While recent KV cache compression methods identify and load important tokens, they focus predominantly on input contexts and fail to address the cumulative attention erro…

Second-order Optimization under Heavy-Tailed Noise: Hessian Clipping and Sample Complexity Limits
Abdurakhmon Sadiev, Peter Richt\'arik, Ilyas Fatkhullin
https://arxiv.org/abs/2510.10690

Second-order Optimization under Heavy-Tailed Noise: Hessian Clipping and Sample Complexity Limits
Heavy-tailed noise is pervasive in modern machine learning applications, arising from data heterogeneity, outliers, and non-stationary stochastic environments. While second-order methods can significantly accelerate convergence in light-tailed or bounded-noise settings, such algorithms are often brittle and lack guarantees under heavy-tailed noise -- precisely the regimes where robustness is most critical. In this work, we take a first step toward a theoretical understanding of second-order opt…

Automatic Failure Attribution and Critical Step Prediction Method for Multi-Agent Systems Based on Causal Inference
Guoqing Ma, Jia Zhu, Hanghui Guo, Weijie Shi, Jiawei Shen, Jingjiang Liu, Yidan Liang
https://arxiv.org/abs/2509.08682

Automatic Failure Attribution and Critical Step Prediction Method for Multi-Agent Systems Based on Causal Inference
Multi-agent systems (MAS) are critical for automating complex tasks, yet their practical deployment is severely hampered by the challenge of failure attribution. Current diagnostic tools, which rely on statistical correlations, are fundamentally inadequate; on challenging benchmarks like Who\&When, state-of-the-art methods achieve less than 15\% accuracy in locating the root-cause step of a failure. To address this critical gap, we introduce the first failure attribution framework for MAS groun…

Can We Reliably Rank Model Performance across Domains without Labeled Data?
Veronica Rammouz, Aaron Gonzalez, Carlos Cruzportillo, Adrian Tan, Nicole Beebe, Anthony Rios
https://arxiv.org/abs/2510.09519

Can We Reliably Rank Model Performance across Domains without Labeled Data?
Estimating model performance without labels is an important goal for understanding how NLP models generalize. While prior work has proposed measures based on dataset similarity or predicted correctness, it remains unclear when these estimates produce reliable performance rankings across domains. In this paper, we analyze the factors that affect ranking reliability using a two-step evaluation setup with four base classifiers and several large language models as error predictors. Experiments on t…

GeoLaux: A Benchmark for Evaluating MLLMs' Geometry Performance on Long-Step Problems Requiring Auxiliary Lines
Yumeng Fu, Jiayin Zhu, Lingling Zhang, Bo Zhao, Shaoxuan Ma, Yushun Zhang, Yanrui Wu, Wenjun Wu
https://arxiv.org/abs/2508.06226

GeoLaux: A Benchmark for Evaluating MLLMs' Geometry Performance on Long-Step Problems Requiring Auxiliary Lines
Geometry problem solving (GPS) requires models to master diagram comprehension, logical reasoning, knowledge application, numerical computation, and auxiliary line construction. This presents a significant challenge for Multimodal Large Language Models (MLLMs). However, existing benchmarks for evaluating MLLM geometry skills overlook auxiliary line construction and lack fine-grained process evaluation, making them insufficient for assessing MLLMs' long-step reasoning abilities. To bridge these …

Harnessing Uncertainty: Entropy-Modulated Policy Gradients for Long-Horizon LLM Agents
Jiawei Wang, Jiacai Liu, Yuqian Fu, Yingru Li, Xintao Wang, Yuan Lin, Yu Yue, Lin Zhang, Yang Wang, Ke Wang
https://arxiv.org/abs/2509.09265

Harnessing Uncertainty: Entropy-Modulated Policy Gradients for Long-Horizon LLM Agents
In long-horizon tasks, recent agents based on Large Language Models (LLMs) face a significant challenge that sparse, outcome-based rewards make it difficult to assign credit to intermediate steps. Previous methods mainly focus on creating dense reward signals to guide learning, either through traditional reinforcement learning techniques like inverse reinforcement learning or by using Process Reward Models for step-by-step feedback. In this paper, we identify a fundamental problem in the learni…

Efficient Trie-based Biasing using K-step Prediction for Rare Word Recognition
Chin Yuen Kwok, Jia Qi yip
https://arxiv.org/abs/2509.09196 https://arxiv.or…

Efficient Trie-based Biasing using K-step Prediction for Rare Word Recognition
Contextual biasing improves rare word recognition of ASR models by prioritizing the output of rare words during decoding. A common approach is Trie-based biasing, which gives "bonus scores" to partial hypothesis (e.g. "Bon") that may lead to the generation of the rare word (e.g. "Bonham"). If the full word ("Bonham") isn't ultimately recognized, the system revokes those earlier bonuses. This revocation is limited to beam search and is computationally expensive, particularly for models with larg…

TORSO: Template-Oriented Reasoning Towards General Tasks
Minhyuk Kim, Seungyoon Lee, Heuiseok Lim
https://arxiv.org/abs/2509.09448 https://arxiv.org/pdf/25…

TORSO: Template-Oriented Reasoning Towards General Tasks
The approaches that guide Large Language Models (LLMs) to emulate human reasoning during response generation have emerged as an effective method for enabling them to solve complex problems in a step-by-step manner, thereby achieving superior performance. However, most existing approaches using few-shot prompts to generate responses heavily depend on the provided examples, limiting the utilization of the model's inherent reasoning capabilities. Moreover, constructing task-specific few-shot promp…

Jupiter: Enhancing LLM Data Analysis Capabilities via Notebook and Inference-Time Value-Guided Search
Shuocheng Li, Yihao Liu, Silin Du, Wenxuan Zeng, Zhe Xu, Mengyu Zhou, Yeye He, Haoyu Dong, Shi Han, Dongmei Zhang
https://arxiv.org/abs/2509.09245

Jupiter: Enhancing LLM Data Analysis Capabilities via Notebook and Inference-Time Value-Guided Search
Large language models (LLMs) have shown great promise in automating data science workflows, but existing models still struggle with multi-step reasoning and tool use, which limits their effectiveness on complex data analysis tasks. To address this, we propose a scalable pipeline that extracts high-quality, tool-based data analysis tasks and their executable multi-step solutions from real-world Jupyter notebooks and associated data files. Using this pipeline, we introduce NbQA, a large-scale dat…