Tootfinder

Fence off Anomaly Interference: Cross-Domain Distillation for Fully Unsupervised Anomaly Detection
Xinyue Liu, Jianyuan Wang, Biao Leng, Shuo Zhang
https://arxiv.org/abs/2508.18007

Fence off Anomaly Interference: Cross-Domain Distillation for Fully Unsupervised Anomaly Detection
Fully Unsupervised Anomaly Detection (FUAD) is a practical extension of Unsupervised Anomaly Detection (UAD), aiming to detect anomalies without any labels even when the training set may contain anomalous samples. To achieve FUAD, we pioneer the introduction of Knowledge Distillation (KD) paradigm based on teacher-student framework into the FUAD setting. However, due to the presence of anomalies in the training data, traditional KD methods risk enabling the student to learn the teacher's repres…

from my link log —
Teacher effects on student achievement and height: a cautionary tale.
https://www.nber.org/papers/w26480
saved 2019-11-27 https://…

Teacher Effects on Student Achievement and Height: A Cautionary Tale
Founded in 1920, the NBER is a private, non-profit, non-partisan organization dedicated to conducting economic research and to disseminating research findings among academics, public policy makers, and business professionals.

Who would have guessed.
Big Tech is paying millions to train teachers on AI, in a push to bring chatbots into classrooms
https://apnews.com/article/artificial-intelligence-teacher-union-microsoft-f7554b6550fb90519dd8129acac8e2…

Parameter-Free Logit Distillation via Sorting Mechanism
Stephen Ekaputra Limantoro
https://arxiv.org/abs/2508.16544 https://arxiv.org/pdf/2508.16544…

Parameter-Free Logit Distillation via Sorting Mechanism
Knowledge distillation (KD) aims to distill the knowledge from the teacher (larger) to the student (smaller) model via soft-label for the efficient neural network. In general, the performance of a model is determined by accuracy, which is measured with labels. However, existing KD approaches usually use the teacher with its original distribution, neglecting the potential of incorrect prediction. This may contradict the motivation of hard-label learning through cross-entropy loss, which may lead…

#ShamelesslyStolenFromSomewhereElseOnTheInternetHonestlyICantKeepTrackOfThisStuffAnymore

So today Dad & I visited a care home Mum will likely move to this week. The asst manager asked what I do for a living, so I told her environmental archaeology, and work for the family business. She said that she'd thought I was a teacher or solicitor by the way I spoke, and I've had that sort of thing said to me a few times. We'd been conveying medical info, clarity is important, so I think my @…

Teacher goes viral with classroom song about sexual consent | Lifestyle | Independent TV
https://www.independent.co.uk/tv/lifestyle/teacher-viral-song-consent-video-b2812508.html

CYCLE-INSTRUCT: Fully Seed-Free Instruction Tuning via Dual Self-Training and Cycle Consistency
Zhanming Shen, Hao Chen, Yulei Tang, Shaolin Zhu, Wentao Ye, Xiaomeng Hu, Haobo Wang, Gang Chen, Junbo Zhao
https://arxiv.org/abs/2508.16100

CYCLE-INSTRUCT: Fully Seed-Free Instruction Tuning via Dual Self-Training and Cycle Consistency
Instruction tuning is vital for aligning large language models (LLMs) with human intent, but current methods typically rely on costly human-annotated seed data or powerful external teacher models. While instruction back-translation techniques reduce this dependency, they remain fundamentally tethered to an initial seed set, which limits full automation, introduces biases, and can lead to inefficient use of unlabeled corpora. In this paper, we propose Cycle-Instruct, a novel framework that achie…

This conservative county lost teachers after administration cut key grant (Ben Brasch/Washington Post)
https://www.washingtonpost.com/education/2025/08/23/trump-cuts-teacher-grants-nc/
http://www.memeorandum.com/250823/p44#a250823p44

Additions with unique digits: A tale of puzzling and AI
The other day, Markus Krötzsch and I were catching up when one of his kids came in. She told us about a puzzle her math teacher gave her:
How many integer sums are there where the equation uses each digit at most once?
The simplest example is 1 2=3, but 12 47=59 also works. On the other hand, 1 9=10 isn’t a solution because the digit ‘1’ appears twice (in the first number and the sum).
1/19

Microsoft, OpenAI, and Anthropic are providing millions to the American Federation of Teachers to build AI training hubs aimed at educating 400,000 teachers (Jocelyn Gecker/Associated Press)
https://apnews.com/article/artificial-int…

Microsoft and OpenAI invest millions in AI training for teachers
The country’s largest teachers unions are teaming up with tech companies like Microsoft and OpenAI to help educators build literacy in artificial intelligence. The tech giants are paying for training for teachers to apply AI skills in the classroom. For the companies, the partnerships are also opportunities to make inroads into schools and win over students in the race for AI dominance. American Federation of Teachers president Randi Weingarten says skepticism guided her negotiations, but the…

DISPATCH: Distilling Selective Patches for Speech Enhancement
Dohwan Kim, Jung-Woo Choi
https://arxiv.org/abs/2509.15922 https://arxiv.org/pdf/2509.15922…

DISPATCH: Distilling Selective Patches for Speech Enhancement
In speech enhancement, knowledge distillation (KD) compresses models by transferring a high-capacity teacher's knowledge to a compact student. However, conventional KD methods train the student to mimic the teacher's output entirely, which forces the student to imitate the regions where the teacher performs poorly and to apply distillation to the regions where the student already performs well, which yields only marginal gains. We propose Distilling Selective Patches (DISPatch), a KD framework …

Having a studio near a piano teacher's studio makes me realise how most kids don't practice this instrument at all, just come and struggle through the same boring tunes every week.
Please don't make your children learn an instrument they don't actual enjoy playing, it's a counter-productive waste of time.
All that stuff about classical music making kids more intelligent is pure nonsense, either experimental results don't replicate or replicate just as we…

Alabama teacher calls state’s new law banning phones in class ‘magic’ | The Independent
https://www.independent.co.uk/news/world/americas/alabama-teacher-cell-phone-ban-law-b2810870.html

Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision
Dulhan Jayalath, Shashwat Goel, Thomas Foster, Parag Jain, Suchin Gururangan, Cheng Zhang, Anirudh Goyal, Alan Schelten
https://arxiv.org/abs/2509.14234

Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision
Where do learning signals come from when there is no ground truth in post-training? We propose turning exploration into supervision through Compute as Teacher (CaT), which converts the model's own exploration at inference-time into reference-free supervision by synthesizing a single reference from a group of parallel rollouts and then optimizing toward it. Concretely, the current policy produces a group of rollouts; a frozen anchor (the initial policy) reconciles omissions and contradictions to…

TOM: An Open-Source Tongue Segmentation Method with Multi-Teacher Distillation and Task-Specific Data Augmentation
Jiacheng Xie, Ziyang Zhang, Biplab Poudel, Congyu Guo, Yang Yu, Guanghui An, Xiaoting Tang, Lening Zhao, Chunhui Xu, Dong Xu
https://arxiv.org/abs/2508.14932

TOM: An Open-Source Tongue Segmentation Method with Multi-Teacher Distillation and Task-Specific Data Augmentation
Tongue imaging serves as a valuable diagnostic tool, particularly in Traditional Chinese Medicine (TCM). The quality of tongue surface segmentation significantly affects the accuracy of tongue image classification and subsequent diagnosis in intelligent tongue diagnosis systems. However, existing research on tongue image segmentation faces notable limitations, and there is a lack of robust and user-friendly segmentation tools. This paper proposes a tongue image segmentation model (TOM) based on…

An Empirical Study of Knowledge Distillation for Code Understanding Tasks
Ruiqi Wang, Zezhou Yang, Cuiyun Gao, Xin Xia, Qing Liao
https://arxiv.org/abs/2508.15423 https://

An Empirical Study of Knowledge Distillation for Code Understanding Tasks
Pre-trained language models (PLMs) have emerged as powerful tools for code understanding. However, deploying these PLMs in large-scale applications faces practical challenges due to their computational intensity and inference latency. Knowledge distillation (KD), a promising model compression and acceleration technique, addresses these limitations by transferring knowledge from large teacher models to compact student models, enabling efficient inference while preserving most of the teacher mode…

FACET:Teacher-Centred LLM-Based Multi-Agent Systems-Towards Personalized Educational Worksheets
Jana Gonnermann-M\"uller, Jennifer Haase, Konstantin Fackeldey, Sebastian Pokutta
https://arxiv.org/abs/2508.11401

FACET:Teacher-Centred LLM-Based Multi-Agent Systems-Towards Personalized Educational Worksheets
The increasing heterogeneity of student populations poses significant challenges for teachers, particularly in mathematics education, where cognitive, motivational, and emotional differences strongly influence learning outcomes. While AI-driven personalization tools have emerged, most remain performance-focused, offering limited support for teachers and neglecting broader pedagogical needs. This paper presents the FACET framework, a teacher-facing, large language model (LLM)-based multi-agent s…

This is comparable to the rise of totalitarian regimes in the 1920s and 1930s,
when Mussolini said even a teacher of mathematics must be a fascist.
Now all who do not take positions on American politics, policies and history that comply with the administration’s views
are in danger of being denied funding, subjected to lawsuits, and derided by the White House in ways that can inspire violent private attacks.
All this has precedents, but not in America’s peacetime hist…

LEAF: Knowledge Distillation of Text Embedding Models with Teacher-Aligned Representations
Robin Vujanic, Thomas Rueckstiess
https://arxiv.org/abs/2509.12539 https://

LEAF: Knowledge Distillation of Text Embedding Models with Teacher-Aligned Representations
We present LEAF ("Lightweight Embedding Alignment Framework"), a knowledge distillation framework for text embedding models. A key distinguishing feature is that our distilled leaf models are aligned to their teacher. In the context of information retrieval, this allows for flexible asymmetric architectures where documents are encoded with the larger teacher model, while queries can be served with the smaller leaf models. We also show that leaf models automatically inherit MRL and robustness to…

Joe Milton plays teacher to Dak Prescott with viral dance lesson https://www.si.com/nfl/cowboys/news/joe-milton-plays-teacher-dak-prescott-viral-dance-lesson-dallas-cowboys-practice

More Texas districts adopt four-day school weeks | The Texas Tribune #Texas

Four-day school weeks are on the rise as Texas districts look for teacher perks on a tight budget
More than 500 schools are using four-day school weeks, a dramatic increase from two years ago driven by tight budgets and fierce competition for educators amid a statewide teacher shortage.

🇺🇦 #NowPlaying on KEXP's #Early
Slick Rick:
🎵 Teacher, Teacher
#SlickRick
https://djbanekroll.bandcamp.com/track/9-slick-rick-teacher-teacher-drake-305-my-city-mashup
https://open.spotify.com/track/77S5EqAq3hkVbAepvHkwLy

@… What an excellent teacher. 😊

"Administrators don’t have the guts to ban cellphones but expect teachers to spend every day swinging wildly at the slithering AI tech we’re supposed to embrace"
https://sunny.garden/@himantra/115039571729383600

Frag mich ob die Analogie zu "KI"-Lernen funkioniert:
Whats the point of "AI" Feedback, if I still have to pay a teacher to fix it?
Wenn wir annehmen, dass Lernen ein ähnlich komplexer Vorgang wie programmieren ist, dann können wir das übertragen. Vielleicht ist ja "KI"-Lernen wenig mehr als ein Rollenspiel für Leute, die sich wie Lehrende fühlen wollen, ohne die eigentliche Arbeit zu machen.
Bringt uns der Begriff vibe learning weiter?

tante (@tante@tldr.nettime.org)
Angehängt: 1 Bild I think the Vibecoding reddit has accidentally stumbled on the best description of vibecoding: It's "roleplay for guys [it is always guys] who want to feel like hackers without doing the hard part". (Source: https://www.reddit.com/r/vibecoding/comments/1mu6t8z/whats_the_point_of_vibe_coding_if_i_still_have_to/ ) #ai #vibecoding

I'm saddened to hear of the passing of my fellow Wesley UMC congregant Bill Platt: https://www.legacy.com/us/obituaries/concordmonitor/name/william-platt-obituary?id=59537900
A Sunday School teacher, youth group leader, and general mentor figure, Bill was just an all-around great guy. The same age as my parents, too! His family and my family were relatively close; I hope they're doing okay...

Exclusive: Oklahoma to begin controversial test to weed out 'woke' teacher applicants today (Sunlen Serfaty/CNN)
https://cnn.com/politics/prageru-oklahoma-woke-teacher-test
http://www.memeorandum.com/250815/p59#a250815p59

Was talking to an acquaintance at church yesterday and she mentioned that things are not good with her job. She’s a secondary school biology teacher. She said more by what she didn’t say. She just clammed up and looked lost. Next time I see her, I want to tell her that it doesn’t have to be that way. That she has transferable skills and can find work that pays the bills without destroying her soul. That there are other places that will value and support her.

Explainable Knowledge Distillation for Efficient Medical Image Classification
Aqib Nazir Mir, Danish Raza Rizvi
https://arxiv.org/abs/2508.15251 https://ar…

Explainable Knowledge Distillation for Efficient Medical Image Classification
This study comprehensively explores knowledge distillation frameworks for COVID-19 and lung cancer classification using chest X-ray (CXR) images. We employ high-capacity teacher models, including VGG19 and lightweight Vision Transformers (Visformer-S and AutoFormer-V2-T), to guide the training of a compact, hardware-aware student model derived from the OFA-595 supernet. Our approach leverages hybrid supervision, combining ground-truth labels with teacher models' soft targets to balance accuracy…

Teaching Audio Models to Reason: A Unified Framework for Source- and Layer-wise Distillation
Runyan Yang, Yuke Si, Yingying Gao, Junlan Feng, Chao Deng, Shilei Zhang
https://arxiv.org/abs/2509.18579

Teaching Audio Models to Reason: A Unified Framework for Source- and Layer-wise Distillation
While large audio language models excel at tasks like ASR and emotion recognition, they still struggle with complex reasoning due to the modality gap between audio and text as well as the lack of structured intermediate supervision. To address this, we propose a unified knowledge distillation framework to transfer reasoning capabilities from a high-capacity textual teacher model to a student audio models while preserving its acoustic competence. Our method introduces two key dimensions: source-…

This is comparable to the rise of totalitarian regimes in the 1920s and 1930s,
when Mussolini said even a teacher of mathematics must be a fascist.
Now all who do not take positions on American politics, policies and history that comply with the administration’s views
are in danger of being denied funding, subjected to lawsuits, and derided by the White House in ways that can inspire violent private attacks.
All this has precedents, but not in America’s peacetime hist…

Opinion | The Mind-Boggling Intrusiveness of Donald J. Trump
The president is driving the tentacles of the federal government deep into the nation’s economy, culture and legal system.

Teacher-Guided Pseudo Supervision and Cross-Modal Alignment for Audio-Visual Video Parsing
Yaru Chen, Ruohao Guo, Liting Gao, Yang Xiang, Qingyu Luo, Zhenbo Li, Wenwu Wang
https://arxiv.org/abs/2509.14097

Teacher-Guided Pseudo Supervision and Cross-Modal Alignment for Audio-Visual Video Parsing
Weakly-supervised audio-visual video parsing (AVVP) seeks to detect audible, visible, and audio-visual events without temporal annotations. Previous work has emphasized refining global predictions through contrastive or collaborative learning, but neglected stable segment-level supervision and class-aware cross-modal alignment. To address this, we propose two strategies: (1) an exponential moving average (EMA)-guided pseudo supervision framework that generates reliable segment-level masks via a…

ClassMind: Scaling Classroom Observation and Instructional Feedback with Multimodal AI
Ao Qu, Yuxi Wen, Jiayi Zhang, Yunge Wen, Yibo Zhao, Alok Prakash, Andr\'es F. Salazar-G\'omez, Paul Pu Liang, Jinhua Zhao
https://arxiv.org/abs/2509.18020

ClassMind: Scaling Classroom Observation and Instructional Feedback with Multimodal AI
Classroom observation -- one of the most effective methods for teacher development -- remains limited due to high costs and a shortage of expert coaches. We present ClassMind, an AI-driven classroom observation system that integrates generative AI and multimodal learning to analyze classroom artifacts (e.g., class recordings) and deliver timely, personalized feedback aligned with pedagogical practices. At its core is AVA-Align, an agent framework that analyzes long classroom video recordings to…

Training thermodynamic computers by gradient descent
Stephen Whitelam
https://arxiv.org/abs/2509.15324 https://arxiv.org/pdf/2509.15324

Training thermodynamic computers by gradient descent
We show how to adjust the parameters of a thermodynamic computer by gradient descent in order to perform a desired computation at a specified observation time. Within a digital simulation of a thermodynamic computer, training proceeds by maximizing the probability with which the computer would generate an idealized dynamical trajectory. The idealized trajectory is designed to reproduce the activations of a neural network trained to perform the desired computation. This teacher-student scheme re…

Took an RSA test today and the handbooks weren't there, teacher filled the time with conversations about ChatGPT "being open-source" and that the under-16s ban is gonna be fine, I think I deserve the license after going through 6 and a half hours of that.

In elementary school, one of my classmates was hit by a car on his bicycle. He must have been about 9. His head went through the windshield.
He was fine.
The windshield was not fine. His helmet was not fine. The teacher had him bring in his mangled helmet to show the whole class. Nobody had to tell us he’d be dead if it weren’t for the helmet. It was obvious.
I still remember this vividly 40 years later.
I wear a helmet.
Re this from @…:
https://hachyderm.io/@thomasfuchs/115364452251790863

Feedback Indicators: The Alignment between Llama and a Teacher in Language Learning
Sylvio R\"udian, Yassin Elsir, Marvin Kretschmer, Sabine Cayrou, Niels Pinkwart
https://arxiv.org/abs/2508.11364

Feedback Indicators: The Alignment between Llama and a Teacher in Language Learning
Automated feedback generation has the potential to enhance students' learning progress by providing timely and targeted feedback. Moreover, it can assist teachers in optimizing their time, allowing them to focus on more strategic and personalized aspects of teaching. To generate high-quality, information-rich formative feedback, it is essential first to extract relevant indicators, as these serve as the foundation upon which the feedback is constructed. Teachers often employ feedback criteria g…

This doesn't look anything like a cell, huh? Well show me your effin' cell model then, all-knowing middle school teacher.

'...in 1981, an English teacher reportedly wrote to the headteacher asking him to reconsider his decision to appoint him [Garage] a prefect, citing reports of fascist views. Another teacher allegedly said that on a Combined Cadet Force camp organised by the college, Farage and others “marched through a quiet Sussex village very late at night shouting Hitler Youth songs".'
Nigel Farage's most controversial moments: from 'racist' comments to a plane crash | The St…

Nigel Farage's most controversial moments: from 'racist' comments to a plane crash
Dubbed 'the man who broke Britain' for his role in Brexit, and viewed by many as a far-right bigot, the Reform UK leader has been in and out of the headlines over the years

Adversarial Distilled Retrieval-Augmented Guarding Model for Online Malicious Intent Detection
Yihao Guo, Haocheng Bian, Liutong Zhou, Ze Wang, Zhaoyi Zhang, Francois Kawala, Milan Dean, Ian Fischer, Yuantao Peng, Noyan Tokgozoglu, Ivan Barrientos, Riyaaz Shaik, Rachel Li, Chandru Venkataraman, Reza Shifteh Far, Moses Pawar, Venkat Sundaranatha, Michael Xu, Frank Chu

Adversarial Distilled Retrieval-Augmented Guarding Model for Online Malicious Intent Detection
With the deployment of Large Language Models (LLMs) in interactive applications, online malicious intent detection has become increasingly critical. However, existing approaches fall short of handling diverse and complex user queries in real time. To address these challenges, we introduce ADRAG (Adversarial Distilled Retrieval-Augmented Guard), a two-stage framework for robust and efficient online malicious intent detection. In the training stage, a high-capacity teacher model is trained on adv…

Membership and Memorization in LLM Knowledge Distillation
Ziqi Zhang, Ali Shahin Shamsabadi, Hanxiao Lu, Yifeng Cai, Hamed Haddadi
https://arxiv.org/abs/2508.07054 https://

Membership and Memorization in LLM Knowledge Distillation
Recent advances in Knowledge Distillation (KD) aim to mitigate the high computational demands of Large Language Models (LLMs) by transferring knowledge from a large ''teacher'' to a smaller ''student'' model. However, students may inherit the teacher's privacy when the teacher is trained on private data. In this work, we systematically characterize and investigate membership and memorization privacy risks inherent in six LLM KD techniques. Using instruction-tuning settings that span seven NLP t…

This is the dreck they want to foist off on WA schools???
In Copilot In Excel Demo, AI Told Teacher a 27% Exam Score Is of No Concern - Slashdot
https://slashdot.org/story/25/10/12/0328228/in-copilot-in-excel-demo-ai-told-teacher-a-27-exam-score-is-of-no-concern

Examining Gender Differences in the Impact of Instructional Clarity on Year 9 Students' Mathematics Interest in New Zealand: A Multi-Group Comparison of 2019 TIMSS Data
Huayu Gao, Tanya Evans, Gavin T. L. Brown
https://arxiv.org/abs/2508.13502

Examining Gender Differences in the Impact of Instructional Clarity on Year 9 Students' Mathematics Interest in New Zealand: A Multi-Group Comparison of 2019 TIMSS Data
Grounded in Social Cognitive Career Theory (SCCT), this study explores how teacher instruction clarity impacts Year 9 students' mathematics interest through the mediating roles of confidence and value, with a focus on gender differences. Structural equation modelling applied to the 2019 New Zealand TIMSS dataset (2,594 girls and 2,651 boys) supports the SCCT framework. While confidence and value have similar mediating effects for boys, confidence plays a more pivotal role for girls. Gender diff…

Day 20: bell hooks.
Despite having decided to continue to 30, number 20 feels important, and hooks gets the spot in part because I haven't yet included a non-fiction feminist author, which feels like an obvious thing to include on such a list. The one category of author being bumped out of the first 20 here is anime writers, but I'll follow up with one of them, along with more academics and mangaka who I've been itching to include.
In any case, hooks is absolutely legendary as a feminist writer for good reason, and as a teacher I've especially appreciated her writing on pedagogy like "Teaching to Transgress: Education as the Practice of Freedom" and "Teaching Critical Thinking: Practical Wisdom". These have challenged me to teach at a higher level, and while I'm not sure I've completely succeeded, they're important to me. They also pair well with Paolo Friere's "Pedagogy of the Oppressed", but hooks always seems to be focused on very practical advice and it's incredibly direct in her writing, even though her advice isn't always straightforward to implement. In fact, that's one of the things I value about her writing: when the truth is complicated or the real work is messy interpersonal relationships that need to be negotiated with each student, she's not afraid to say so and give good advice for navigating those waters instead of trying to dispense simple-seeming platitudes or formulas for success that paper over the deeper issues. Her concern has always been truth, rather than simplicity or audience comfort and the popularity it might seem to entail, which I think is part of why her legacy endures so well.
#20AuthorsNoMen
#30AuthorsNoMen

@… In the end, it’s just another warmed-up version of the age-old vision of the robo-teacher that’ll make learning so much more efficient—and cheaper, too!

@… In the end, it’s just another warmed-up version of the age-old vision of the robo-teacher that’ll make learning so much more efficient—and cheaper, too!

Retired high school teacher, and former #PurdueFortWayne 🐘 Limited-term Lecturer and Visiting Instructor Bob Lovell is featured in this local #PBS Arts IN Focus video about his mathematical quilt-making hobby.
12:27 YouTube Video

Embedding the Teacher: Distilling vLLM Preferences for Scalable Image Retrieval
Eric He, Akash Gupta, Adian Liusie, Vatsal Raina, Piotr Molenda, Shirom Chabra, Vyas Raina
https://arxiv.org/abs/2510.12014

Embedding the Teacher: Distilling vLLM Preferences for Scalable Image Retrieval
Text--image retrieval is necessary for applications such as product recommendation. Embedding-based approaches like CLIP enable efficient large-scale retrieval via vector similarity search, but they are primarily trained on literal caption-like text--image pairs and often fail to capture abstract or persona-driven attributes common in product recommendation applications (e.g., ``a gift for a mother who loves gardening''). In contrast, state-of-the-art vision--language models (vLLMs) can align t…

LearnLens: An AI-Enhanced Dashboard to Support Teachers in Open-Ended Classrooms
Namrata Srivastava, Shruti Jain, Clayton Cohn, Naveeduddin Mohammed, Umesh Timalsina, Gautam Biswas
https://arxiv.org/abs/2509.10582

LearnLens: An AI-Enhanced Dashboard to Support Teachers in Open-Ended Classrooms
Exploratory learning environments (ELEs), such as simulation-based platforms and open-ended science curricula, promote hands-on exploration and problem-solving but make it difficult for teachers to gain timely insights into students' conceptual understanding. This paper presents LearnLens, a generative AI (GenAI)-enhanced teacher-facing dashboard designed to support problem-based instruction in middle school science. LearnLens processes students' open-ended responses from digital assessments to…

EMNLP: Educator-role Moral and Normative Large Language Models Profiling
Yilin Jiang, Mingzi Zhang, Sheng Jin, Zengyi Yu, Xiangjie Kong, Binghao Tu
https://arxiv.org/abs/2508.15250

EMNLP: Educator-role Moral and Normative Large Language Models Profiling
Simulating Professions (SP) enables Large Language Models (LLMs) to emulate professional roles. However, comprehensive psychological and ethical evaluation in these contexts remains lacking. This paper introduces EMNLP, an Educator-role Moral and Normative LLMs Profiling framework for personality profiling, moral development stage measurement, and ethical risk under soft prompt injection. EMNLP extends existing scales and constructs 88 teacher-specific moral dilemmas, enabling profession-orient…

🇺🇦 #NowPlaying on KEXP's #Early
Van Halen:
🎵 Hot for Teacher
#VanHalen
https://cigaweededits.bandcamp.com/track/van-halen-hot-for-teacher-dj-l-h-d-m-edit
https://open.spotify.com/track/6QDbGdbJ57Mtkflsg42WV5

Incorporating priors in learning: a random matrix study under a teacher-student framework
Malik Tiomoko, Ekkehard Schnoor
https://arxiv.org/abs/2509.22124 https://

Incorporating priors in learning: a random matrix study under a teacher-student framework
Regularized linear regression is central to machine learning, yet its high-dimensional behavior with informative priors remains poorly understood. We provide the first exact asymptotic characterization of training and test risks for maximum a posteriori (MAP) regression with Gaussian priors centered at a domain-informed initialization. Our framework unifies ridge regression, least squares, and prior-informed estimators, and -- using random matrix theory -- yields closed-form risk formulas that …

This is a bit ancient (2013?) but some neat ideas here... I like the sliding potentiometer to sense when the brake cable is pulled.
#biking

jdeboi.com
Jenna deBoisblanc is a creative coder, teacher, and maker.

Intercultural perspective in teacher training: Astronomy as an entry point
Mayra Manente, Geraldine Chadwick, Alejandro Gangui
https://arxiv.org/abs/2509.00254 https://

Intercultural perspective in teacher training: Astronomy as an entry point
With the aim of incorporating an intercultural perspective into Physics teacher training programs, we ask ourselves which current disciplinary contents could begin to be addressed from this perspective. We started out by exploring the current state of the field exclusively in relation to intercultural science education, focusing on direct relationships with Physics. We present here a synthesis of publications from the past five years that we have found so far. These publications were categorize…

Merge-of-Thought Distillation
Zhanming Shen, Zeyu Qin, Zenan Huang, Hao Chen, Jiaqi Hu, Yihong Zhuang, Guoshan Lu, Gang Chen, Junbo Zhao
https://arxiv.org/abs/2509.08814 https:/…

Merge-of-Thought Distillation
Efficient reasoning distillation for long chain-of-thought (CoT) models is increasingly constrained by the assumption of a single oracle teacher, despite practical availability of multiple candidate teachers and growing CoT corpora. We revisit teacher selection and observe that different students have different "best teachers," and even for the same student the best teacher can vary across datasets. Therefore, to unify multiple teachers' reasoning abilities into student with overcoming conflict…

Oklahoma to test teachers from New York, California to guard against 'radical leftist ideology' (Associated Press)
https://apnews.com/article/oklahoma-teacher-test-ideology-california-new-york-1d08d0ef4e6e3dc71d8537601d8455e9
http://www.memeorandum.com/250818/p152#a250818p152

Diverse and Adaptive Behavior Curriculum for Autonomous Driving: A Student-Teacher Framework with Multi-Agent RL
Ahmed Abouelazm, Johannes Ratz, Philip Sch\"orner, J. Marius Z\"ollner
https://arxiv.org/abs/2507.19146

Diverse and Adaptive Behavior Curriculum for Autonomous Driving: A Student-Teacher Framework with Multi-Agent RL
Autonomous driving faces challenges in navigating complex real-world traffic, requiring safe handling of both common and critical scenarios. Reinforcement learning (RL), a prominent method in end-to-end driving, enables agents to learn through trial and error in simulation. However, RL training often relies on rule-based traffic scenarios, limiting generalization. Additionally, current scenario generation methods focus heavily on critical scenarios, neglecting a balance with routine driving beh…

This video reminded me of my 5th grade teacher who mocked me in front of the class for giving the answer "Pasteurization" (attributed, according to her, to a "coincidence") to her question "what things have Louis Pasteur invented?"
(Hint: she was expecting "rabies vaccine" as an answer.)
https://www.

🔊 #NowPlaying on #BBCRadio3:
#TheLebrechtInterview
- Harpsichordist Mahan Esfahani
Norman Lebrecht talks to Mahan Esfahani about choosing harpsichord or piano; about not being part of the Early Music movement; and about her legendary teacher Zuzana Ruzickova.
Relisten now 👇
https://www.bbc.co.uk/programmes/m002gryt

#WritersCoffeeClub Aug 7: Have you always written, or was there a tipping point to get you to start?
I was the reason our teacher set a max page count to the essays and short stories. The stories, the tales, and the fantasies have always been a part of me. But I only recently actually tried to catch anything in ink seriously.

Replaced article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[8/9]:
- More Than One Teacher: Adaptive Multi-Guidance Policy Optimization for Diverse Exploration
Xiaoyang Yuan, Yujuan Ding, Yi Bin, Wenqi Shao, Jinyu Cai, Jingkuan Song, Yang Yang, Heng Tao Shen

Harvard Crimson Student Reflection
Harvard Dismantled the Program That Changed My Life Three weeks ago, HDS announced it would “suspend” RCPI amidst budget cuts and accusations of one-sidedness. HDS also let go of Hilary Rantisi, a dedicated teacher, scholar, and mentor — and the only Palestinian American employed at HDS. In January, Diane L. Moore, the Associate Dean of the Religion and Public Life program at HDS — which houses RCPI — abruptly departed a semester…

Self-Supervised Stereo Matching with Multi-Baseline Contrastive Learning
Peng Xu, Zhiyu Xiang, Jingyun Fu, Tianyu Pu, Kai Wang, Chaojie Ji, Tingming Bai, Eryun Liu
https://arxiv.org/abs/2508.10838

Self-Supervised Stereo Matching with Multi-Baseline Contrastive Learning
Current self-supervised stereo matching relies on the photometric consistency assumption, which breaks down in occluded regions due to ill-posed correspondences. To address this issue, we propose BaCon-Stereo, a simple yet effective contrastive learning framework for self-supervised stereo network training in both non-occluded and occluded regions. We adopt a teacher-student paradigm with multi-baseline inputs, in which the stereo pairs fed into the teacher and student share the same reference …

Kid2, after learning who she'll have for a teacher this year: "YYYYYYYYYYEEEEEEEEEEEEESSSSSSSSSSSSSSSSSSS!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
There's nothing like a good teacher.

Karoline Leavitt is my first-grade teacher from hell #BrainDamageAndLies

Once Upon a Time: Interactive Learning for Storytelling with Small Language Models
Jonas Mayer Martins, Ali Hamza Bashir, Muhammad Rehan Khalid, Lisa Beinborn
https://arxiv.org/abs/2509.15714

Once Upon a Time: Interactive Learning for Storytelling with Small Language Models
Children efficiently acquire language not just by listening, but by interacting with others in their social environment. Conversely, large language models are typically trained with next-word prediction on massive amounts of text. Motivated by this contrast, we investigate whether language models can be trained with less data by learning not only from next-word prediction but also from high-level, cognitively inspired feedback. We train a student model to generate stories, which a teacher model…

Adaptive Knowledge Distillation using a Device-Aware Teacher for Low-Complexity Acoustic Scene Classification
Seung Gyu Jeong, Seong Eun Kim
https://arxiv.org/abs/2509.09262 htt…

Adaptive Knowledge Distillation using a Device-Aware Teacher for Low-Complexity Acoustic Scene Classification
In this technical report, we describe our submission for Task 1, Low-Complexity Device-Robust Acoustic Scene Classification, of the DCASE 2025 Challenge. Our work tackles the dual challenges of strict complexity constraints and robust generalization to both seen and unseen devices, while also leveraging the new rule allowing the use of device labels at test time. Our proposed system is based on a knowledge distillation framework where an efficient CP-MobileNet student learns from a compact, spe…

Taught Well Learned Ill: Towards Distillation-conditional Backdoor Attack
Yukun Chen, Boheng Li, Yu Yuan, Leyi Qi, Yiming Li, Tianwei Zhang, Zhan Qin, Kui Ren
https://arxiv.org/abs/2509.23871

Taught Well Learned Ill: Towards Distillation-conditional Backdoor Attack
Knowledge distillation (KD) is a vital technique for deploying deep neural networks (DNNs) on resource-constrained devices by transferring knowledge from large teacher models to lightweight student models. While teacher models from third-party platforms may undergo security verification (\eg, backdoor detection), we uncover a novel and critical threat: distillation-conditional backdoor attacks (DCBAs). DCBA injects dormant and undetectable backdoors into teacher models, which become activated i…

"Coch and Daniel do not call for more neuroscience in classrooms, but for better scientific literacy among teachers, school leaders, and teacher educators. They demonstrate that anyone seeking to apply brain research to enhance teaching must first comprehend what that research can – and cannot – reveal."

The stereotype for this kind of self-narrowing is math/tech people avoiding humanities. That certainly happens, but I see it at least as often in the other direction: people avoiding anything that smells like STEM, telling themselves they can’t, building a whole identity around it. It comes out as this weird kind of shame/pride mix — “I’m not a math person” etc — which tends to hide some kind of earlier trauma: bad teacher, bad curriculum, gender role policing, who knows.

‘Are You Going to Stay Awake?’: Pennsylvania Teacher Tied Black Student to a Board By His Neck for Napping In Class, Lawsuit Says
https://atlantablackstar.com/2025/08/31/pennsylvania-teacher-tied-black-student-to-a-board/

Personalized Inhibition Training with Eye-Tracking: Enhancing Student Learning and Teacher Assessment in Educational Games
Abdul Rehman, Ilona Heldal, Diana Stilwell, Paula Costa Ferreira, Jerry Chun-Wei Lin
https://arxiv.org/abs/2509.08357

Personalized Inhibition Training with Eye-Tracking: Enhancing Student Learning and Teacher Assessment in Educational Games
Eye tracking (ET) can help to understand visual attention and cognitive processes in interactive environments. This study presents a comprehensive eye-tracking analysis framework of the Inhibitory Control Game, named the ReStroop game, which is an educational intervention aimed at improving inhibitory control skills in children through a recycling-themed sorting task, for educational assessment that processes raw gaze data through unified algorithms for fixation detection, performance evaluatio…

Distilled-3DGS:Distilled 3D Gaussian Splatting
Lintao Xiang, Xinkai Chen, Jianhuang Lai, Guangcong Wang
https://arxiv.org/abs/2508.14037 https://arxiv.org/…

Distilled-3DGS:Distilled 3D Gaussian Splatting
3D Gaussian Splatting (3DGS) has exhibited remarkable efficacy in novel view synthesis (NVS). However, it suffers from a significant drawback: achieving high-fidelity rendering typically necessitates a large number of 3D Gaussians, resulting in substantial memory consumption and storage requirements. To address this challenge, we propose the first knowledge distillation framework for 3DGS, featuring various teacher models, including vanilla 3DGS, noise-augmented variants, and dropout-regularize…

From Reasoning LLMs to BERT: A Two-Stage Distillation Framework for Search Relevance
Runze Xia, Yupeng Ji, Yuxi Zhou, Haodong Liu, Teng Zhang, Piji Li
https://arxiv.org/abs/2510.11056

From Reasoning LLMs to BERT: A Two-Stage Distillation Framework for Search Relevance
Query-service relevance prediction in e-commerce search systems faces strict latency requirements that prevent the direct application of Large Language Models (LLMs). To bridge this gap, we propose a two-stage reasoning distillation framework to transfer reasoning capabilities from a powerful teacher LLM to a lightweight, deployment-friendly student model. In the first stage, we address the limitations of general-purpose LLMs by constructing a domain-adapted teacher model. This is achieved thro…

SENSE models: an open source solution for multilingual and multimodal semantic-based tasks
Salima Mdhaffar, Haroun Elleuch, Chaimae Chellaf, Ha Nguyen, Yannick Est\`eve
https://arxiv.org/abs/2509.12093

SENSE models: an open source solution for multilingual and multimodal semantic-based tasks
This paper introduces SENSE (Shared Embedding for N-lingual Speech and tExt), an open-source solution inspired by the SAMU-XLSR framework and conceptually similar to Meta AI's SONAR models. These approaches rely on a teacher-student framework to align a self-supervised speech encoder with the language-agnostic continuous representations of a text encoder at the utterance level. We describe how the original SAMU-XLSR method has been updated by selecting a stronger teacher text model and a better…

Discriminative Feature Feedback with General Teacher Classes
Omri Bar Oz, Tosca Lechner, Sivan Sabato
https://arxiv.org/abs/2510.07245 https://arxiv.org/pd…

Discriminative Feature Feedback with General Teacher Classes
We study the theoretical properties of the interactive learning protocol Discriminative Feature Feedback (DFF) (Dasgupta et al., 2018). The DFF learning protocol uses feedback in the form of discriminative feature explanations. We provide the first systematic study of DFF in a general framework that is comparable to that of classical protocols such as supervised learning and online learning. We study the optimal mistake bound of DFF in the realizable and the non-realizable settings, and obtain …

🇺🇦 #NowPlaying on KEXP's #Early
Prince:
🎵 Teacher, Teacher
#Prince
https://princeazariyah.bandcamp.com/album/h-oly-o-ne-t-eacher

A Random Matrix Perspective of Echo State Networks: From Precise Bias--Variance Characterization to Optimal Regularization
Yessin Moakher, Malik Tiomoko, Cosme Louart, Zhenyu Liao
https://arxiv.org/abs/2509.22011

A Random Matrix Perspective of Echo State Networks: From Precise Bias--Variance Characterization to Optimal Regularization
We present a rigorous asymptotic analysis of Echo State Networks (ESNs) in a teacher student setting with a linear teacher with oracle weights. Leveraging random matrix theory, we derive closed form expressions for the asymptotic bias, variance, and mean-squared error (MSE) as functions of the input statistics, the oracle vector, and the ridge regularization parameter. The analysis reveals two key departures from classical ridge regression: (i) ESNs do not exhibit double descent, and (ii) ESNs …

Cross-Modal Knowledge Distillation with Multi-Level Data Augmentation for Low-Resource Audio-Visual Sound Event Localization and Detection
Qing Wang, Ya Jiang, Hang Chen, Sabato Marco Siniscalchi, Jun Du, Jianqing Gao
https://arxiv.org/abs/2508.12334

Cross-Modal Knowledge Distillation with Multi-Level Data Augmentation for Low-Resource Audio-Visual Sound Event Localization and Detection
This work presents a cross-modal knowledge distillation (CMKD) framework combined with multi-level data augmentation for low-resource audio-visual (AV) sound event localization and detection (SELD). An audio-only SELD model acts as the teacher, transferring knowledge to an AV student model through both output responses and intermediate feature representations. To enhance learning, data augmentation is applied by mixing features randomly selected from multiple network layers and associated loss …

Harvard Crimson Student Reflection
Harvard Dismantled the Program That Changed My Life
Three weeks ago, HDS announced it would “suspend” RCPI amidst budget cuts and accusations of one-sidedness. HDS also let go of Hilary Rantisi, a dedicated teacher, scholar, and mentor — and the only Palestinian American employed at HDS. In January

I know this sounds click-baity, but it's worth reading. Think "hilarious things that kids say" and you're ready to go 🙂 (appeared in my feed because of the education connection)
https://www.upworthy.com/mom-is-completely-embarrassed-ex1

More Than One Teacher: Adaptive Multi-Guidance Policy Optimization for Diverse Exploration
Xiaoyang Yuan, Yujuan Ding, Yi Bin, Wenqi Shao, Jinyu Cai, Jingkuan Song, Yang Yang, Hengtao Shen
https://arxiv.org/abs/2510.02227

More Than One Teacher: Adaptive Multi-Guidance Policy Optimization for Diverse Exploration
Reinforcement Learning with Verifiable Rewards (RLVR) is a promising paradigm for enhancing the reasoning ability in Large Language Models (LLMs). However, prevailing methods primarily rely on self-exploration or a single off-policy teacher to elicit long chain-of-thought (LongCoT) reasoning, which may introduce intrinsic model biases and restrict exploration, ultimately limiting reasoning diversity and performance. Drawing inspiration from multi-teacher strategies in knowledge distillation, we…

Lost in Translation? Vocabulary Alignment for Source-Free Domain Adaptation in Open-Vocabulary Semantic Segmentation
Silvio Mazzucco, Carl Persson, Mattia Segu, Pier Luigi Dovesi, Federico Tombari, Luc Van Gool, Matteo Poggi
https://arxiv.org/abs/2509.15225

Lost in Translation? Vocabulary Alignment for Source-Free Domain Adaptation in Open-Vocabulary Semantic Segmentation
We introduce VocAlign, a novel source-free domain adaptation framework specifically designed for VLMs in open-vocabulary semantic segmentation. Our method adopts a student-teacher paradigm enhanced with a vocabulary alignment strategy, which improves pseudo-label generation by incorporating additional class concepts. To ensure efficiency, we use Low-Rank Adaptation (LoRA) to fine-tune the model, preserving its original capabilities while minimizing computational overhead. In addition, we propos…

GoldMind: A Teacher-Centered Knowledge Management System for Higher Education -- Lessons from Iterative Design
Gloria Fern\'andez-Nieto, Lele Sha, Yuheng Li, Yi-Shan Tsai, Guanliang Chen, Yinwei Wei, Weiqing Wang, Jinchun Wen, Shaveen Singh, Ivan Silva, Yuanfang Li, Dragan Gas\v{e}vi\'c, Zachari Swiecki
https://arxiv.org/abs/2508.0…

GoldMind: A Teacher-Centered Knowledge Management System for Higher Education -- Lessons from Iterative Design
Designing Knowledge Management Systems (KMSs) for higher education requires addressing complex human-technology interactions, especially where staff turnover and changing roles create ongoing challenges for reusing knowledge. While advances in process mining and Generative AI enable new ways of designing features to support knowledge management, existing KMSs often overlook the realities of educators' workflows, leading to low adoption and limited impact. This paper presents findings from a two…

Paired by the Teacher: Turning Unpaired Data into High-Fidelity Pairs for Low-Resource Text Generation
Yen-Ju Lu, Thomas Thebaud, Laureano Moro-Velazquez, Najim Dehak, Jesus Villalba
https://arxiv.org/abs/2509.25144

Paired by the Teacher: Turning Unpaired Data into High-Fidelity Pairs for Low-Resource Text Generation
We present Paired by the Teacher (PbT), a two-stage teacher-student pipeline that synthesizes accurate input-output pairs without human labels or parallel data. In many low-resource natural language generation (NLG) scenarios, practitioners may have only raw outputs, like highlights, recaps, or questions, or only raw inputs, such as articles, dialogues, or paragraphs, but seldom both. This mismatch forces small models to learn from very few examples or rely on costly, broad-scope synthetic exam…

MOCHA: Multi-modal Objects-aware Cross-arcHitecture Alignment
Elena Camuffo, Francesco Barbato, Mete Ozay, Simone Milani, Umberto Michieli
https://arxiv.org/abs/2509.14001 https…

MOCHA: Multi-modal Objects-aware Cross-arcHitecture Alignment
We introduce MOCHA (Multi-modal Objects-aware Cross-arcHitecture Alignment), a knowledge distillation approach that transfers region-level multimodal semantics from a large vision-language teacher (e.g., LLaVa) into a lightweight vision-only object detector student (e.g., YOLO). A translation module maps student features into a joint space, where the training of the student and translator is guided by a dual-objective loss that enforces both local alignment and global relational consistency. Un…

Capturing and Sharing Know-How through Visual Process Representations: A Human-Centred Approach to Teacher Workflows
Gloria Fern\'andez-Nieto, Vanessa Echeverria, Yuheng Li, Yi-Shan Tsai, Lele Sha, Guanliang Chen, Dragan Gasevic, Zachari Swiecki
https://arxiv.org/abs/2508.04357

Capturing and Sharing Know-How through Visual Process Representations: A Human-Centred Approach to Teacher Workflows
Knowledge Management is crucial for capturing and transferring expertise within universities, especially in high staff turnover contexts where expertise loss disrupts teaching. Documenting teachers' workflows is time-intensive and diverts experts from core responsibilities. Sequential Pattern Mining (SPM) leverages log data to identify expert workflows, offering an automated alternative to represent workflows but requiring transformation into intuitive formats for novice educators. This paper i…

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[2/4]:
- Embedding the Teacher: Distilling vLLM Preferences for Scalable Image Retrieval
Eric He, Akash Gupta, Adian Liusie, Vatsal Raina, Piotr Molenda, Shirom Chabra, Vyas Raina

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[2/3]:
- Assistant-Guided Mitigation of Teacher Preference Bias in LLM-as-a-Judge
Zhuo Liu, Moxin Li, Xun Deng, Qifan Wang, Fuli Feng

Improving Adversarial Robustness Through Adaptive Learning-Driven Multi-Teacher Knowledge Distillation
Hayat Ullah, Syed Muhammad Talha Zaidi, Arslan Munir
https://arxiv.org/abs/2507.20996

Improving Adversarial Robustness Through Adaptive Learning-Driven Multi-Teacher Knowledge Distillation
Convolutional neural networks (CNNs) excel in computer vision but are susceptible to adversarial attacks, crafted perturbations designed to mislead predictions. Despite advances in adversarial training, a gap persists between model accuracy and robustness. To mitigate this issue, in this paper, we present a multi-teacher adversarial robustness distillation using an adaptive learning strategy. Specifically, our proposed method first trained multiple clones of a baseline CNN model using an advers…

Knowledge Distillation Detection for Open-weights Models
Qin Shi, Amber Yijia Zheng, Qifan Song, Raymond A. Yeh
https://arxiv.org/abs/2510.02302 https://ar…

Knowledge Distillation Detection for Open-weights Models
We propose the task of knowledge distillation detection, which aims to determine whether a student model has been distilled from a given teacher, under a practical setting where only the student's weights and the teacher's API are available. This problem is motivated by growing concerns about model provenance and unauthorized replication through distillation. To address this task, we introduce a model-agnostic framework that combines data-free input synthesis and statistical score computation f…

DiaCDM: Cognitive Diagnosis in Teacher-Student Dialogues using the Initiation-Response-Evaluation Framework
Rui Jia, Yuang Wei, Ruijia Li, Yuang-Hao Jiang, Xinyu Xie, Yaomin Shen, Min Zhang, Bo Jiang
https://arxiv.org/abs/2509.24821

DiaCDM: Cognitive Diagnosis in Teacher-Student Dialogues using the Initiation-Response-Evaluation Framework
While cognitive diagnosis (CD) effectively assesses students' knowledge mastery from structured test data, applying it to real-world teacher-student dialogues presents two fundamental challenges. Traditional CD models lack a suitable framework for handling dynamic, unstructured dialogues, and it's difficult to accurately extract diagnostic semantics from lengthy dialogues. To overcome these hurdles, we propose DiaCDM, an innovative model. We've adapted the initiation-response-evaluation (IRE) f…

Emerging Patterns of GenAI Use in K-12 Science and Mathematics Education
Lief Esbenshade (Victor), Shawon Sarkar (Victor), Drew Nucci (Victor), Ann Edwards (Victor), Sarah Nielsen (Victor), Joshua M. Rosenberg (Victor), Alex Liu (Victor), Zewei (Victor), Tian, Min Sun, Zachary Zhang, Thomas Han, Yulia Lapicus, Kevin He
https://arxiv.org/…

Emerging Patterns of GenAI Use in K-12 Science and Mathematics Education
In this report, we share findings from a nationally representative survey of US public school math and science teachers, examining current generative AI (GenAI) use, perceptions, constraints, and institutional support. We show trends in math and science teacher adoption of GenAI, including frequency and purpose of use. We describe how teachers use GenAI with students and their beliefs about GenAI's impact on student learning. We share teachers' reporting on the school and district support they …

Hala Technical Report: Building Arabic-Centric Instruction & Translation Models at Scale
Hasan Abed Al Kader Hammoud, Mohammad Zbeeb, Bernard Ghanem
https://arxiv.org/abs/2509.14008

Hala Technical Report: Building Arabic-Centric Instruction & Translation Models at Scale
We present Hala, a family of Arabic-centric instruction and translation models built with our translate-and-tune pipeline. We first compress a strong AR$\leftrightarrow$EN teacher to FP8 (yielding $\sim$2$\times$ higher throughput with no quality loss) and use it to create high-fidelity bilingual supervision. A lightweight language model LFM2-1.2B is then fine-tuned on this data and used to translate high-quality English instruction sets into Arabic, producing a million-scale corpus tailored to…

Circuit Distillation
Somin Wadhwa, Silvio Amir, Byron C. Wallace
https://arxiv.org/abs/2509.25002 https://arxiv.org/pdf/2509.25002

Circuit Distillation
Model distillation typically focuses on behavioral mimicry, where a student model is trained to replicate a teacher's output while treating its internal computations as a black box. In this work we propose an alternative approach: Distilling the underlying computational mechanisms implemented by a teacher model. Specifically, we propose circuit distillation, which introduces an objective to align internal representations between analogous circuit components in teacher and student models. We pro…

Synthetic Data Generation for Emotional Depth Faces: Optimizing Conditional DCGANs via Genetic Algorithms in the Latent Space and Stabilizing Training with Knowledge Distillation
Seyed Muhammad Hossein Mousavi, S. Younes Mirinezhad
https://arxiv.org/abs/2508.09188

Synthetic Data Generation for Emotional Depth Faces: Optimizing Conditional DCGANs via Genetic Algorithms in the Latent Space and Stabilizing Training with Knowledge Distillation
Affective computing faces a major challenge: the lack of high-quality, diverse depth facial datasets for recognizing subtle emotional expressions. We propose a framework for synthetic depth face generation using an optimized GAN with Knowledge Distillation (EMA teacher models) to stabilize training, improve quality, and prevent mode collapse. We also apply Genetic Algorithms to evolve GAN latent vectors based on image statistics, boosting diversity and visual quality for target emotions. The ap…

Uncertainty-aware Cross-training for Semi-supervised Medical Image Segmentation
Kaiwen Huang, Tao Zhou, Huazhu Fu, Yizhe Zhang, Yi Zhou, Xiao-Jun Wu
https://arxiv.org/abs/2508.09014

Uncertainty-aware Cross-training for Semi-supervised Medical Image Segmentation
Semi-supervised learning has gained considerable popularity in medical image segmentation tasks due to its capability to reduce reliance on expert-examined annotations. Several mean-teacher (MT) based semi-supervised methods utilize consistency regularization to effectively leverage valuable information from unlabeled data. However, these methods often heavily rely on the student model and overlook the potential impact of cognitive biases within the model. Furthermore, some methods employ co-tr…

Evaluating LLMs Without Oracle Feedback: Agentic Annotation Evaluation Through Unsupervised Consistency Signals
Cheng Chen, Haiyan Yin, Ivor Tsang
https://arxiv.org/abs/2509.08809

Evaluating LLMs Without Oracle Feedback: Agentic Annotation Evaluation Through Unsupervised Consistency Signals
Large Language Models (LLMs), when paired with prompt-based tasks, have significantly reduced data annotation costs and reliance on human annotators. However, evaluating the quality of their annotations remains challenging in dynamic, unsupervised environments where oracle feedback is scarce and conventional methods fail. To address this challenge, we propose a novel agentic annotation paradigm, where a student model collaborates with a noisy teacher (the LLM) to assess and refine annotation qu…

Parallel-R1: Towards Parallel Thinking via Reinforcement Learning
Tong Zheng, Hongming Zhang, Wenhao Yu, Xiaoyang Wang, Xinyu Yang, Runpeng Dai, Rui Liu, Huiwen Bao, Chengsong Huang, Heng Huang, Dong Yu
https://arxiv.org/abs/2509.07980

Parallel-R1: Towards Parallel Thinking via Reinforcement Learning
Parallel thinking has emerged as a novel approach for enhancing the reasoning capabilities of large language models (LLMs) by exploring multiple reasoning paths concurrently. However, activating such capabilities through training remains challenging, as existing methods predominantly rely on supervised fine-tuning (SFT) over synthetic data, which encourages teacher-forced imitation rather than exploration and generalization. Different from them, we propose \textbf{Parallel-R1}, the first reinfo…