Daily AI Papers — September 6, 2026

1. A Common Measure of Communication for Speech Brain-Computer Interfaces

Authors: Dulhan Jayalath, Benjamin Ballyk, Oiwi Parker Jones arXiv: arxiv.org/abs/2609.02887 Summary: Speech brain-computer interfaces (speech BCIs) translate neural activity into language, offering a path towards restoring speech for people with paralysis and, more broadly, enabling new forms of natural human-computer interaction. Despite this promise, the field lacks a common measure of progress because systems use different datasets, recording methods, types of speech, and vocabularies, so their reported scores are rarely comparable. Trending because: 10 HuggingFace upvotes + offers a common information-theoretic yardstick for comparing speech brain-computer interfaces

2. One Symptom, Three Levers: A Critical Review of On-Policy Self-Distillation

Authors: Justin Robert, Raheel Qader arXiv: arxiv.org/abs/2608.25936 Summary: On-policy distillation trains a language model on its own generations while a teacher scores them token by token. It combines the dense supervision of imitation learning with the on-policy sampling of reinforcement learning. Trending because: 10 HuggingFace upvotes + clarifies the mechanisms and collapse risks of on-policy self-distillation

3. Causal Foundation Models

Authors: Christopher Stith, Hossein Rahmani, Jesse C. Cresswell arXiv: arxiv.org/abs/2609.03003 Summary: Causal inference is the practice of estimating the effect of a treatment or intervention from data. It traditionally requires a bespoke pipeline for every new problem: first proposing a causal mechanism, selecting a compatible estimator, and finally training it. Trending because: 8 HuggingFace upvotes + surveys an emerging foundation-model paradigm for causal inference

4. Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation

Authors: Zhiwei Zhang, Zechen Sun, Fei Zhao, Kang Peng, Bin Liang, Huayu Deng, Yao Hu, Kam-Fai Wong, Mu Chuan arXiv: arxiv.org/abs/2609.02998 Summary: On-policy distillation (OPD) accelerates post-training by providing dense token-level supervision from a frozen teacher on the student’s own rollouts. Vanilla OPD applies this supervision uniformly across prompts, without checking whether the teacher is reliable for each prompt. Trending because: 8 HuggingFace upvotes + gates dense distillation on prompt-level teacher reliability

5. DICS: Exploring Data Intrinsic Consistency for Visual Instruction Selection

Authors: Yuyang Hong, Jinhui Guo, Jiaqi Gu, Lubin Fan, Ruixiang Wang, Kun Ding, Yue Wu, Shiming Xiang, Jieping Ye arXiv: arxiv.org/abs/2608.30209 Summary: Visual instruction tuning is crucial for advancing the vision-language alignment and instruction-following capabilities of Vision-Language Models (VLMs). However, identifying optimal subsets under a fixed ratio constraint from rapidly expanding datasets remains a significant bottleneck. Trending because: 7 HuggingFace upvotes + selects efficient visual-instruction data using intrinsic consistency

6. Language Chain in Alignment: Cross-lingual Ranking Preference Optimization

Authors: Seungyoon Lee, Minhyuk Kim, Jungseob Lee, Heuiseok Lim arXiv: arxiv.org/abs/2608.23149 Summary: The alignment of Large Language Models heavily relies on English-centric high-quality preference data, which often leads to suboptimal performance in other languages. In this paper, we propose Cross-lingual Ranking Preference Optimization~(CRPO), a novel framework that leverages robust preference knowledge from English to facilitate preference alignment in the target language. Trending because: 6 HuggingFace upvotes + transfers preference alignment across languages with richer ranking signals

7. Cross-lingual Functional Vectors for Emotion Detection in Large Language Models

Authors: Jieying Xue, Phuong Minh Nguyen, Minh Le Nguyen, Shogo Okada arXiv: arxiv.org/abs/2608.29613 Summary: Function vectors (FVs) have recently emerged as a promising mechanism for steering the behavior of large language models (LLMs) by injecting task-specific latent direction representations derived from in-context demonstrations. While prior studies have shown that FVs can recover task behavior in structured in-context learning settings, their effectiveness on semantically complex tasks and their ability to generalize across languages remain underexplored. Trending because: 6 HuggingFace upvotes + finds transferable latent steering directions for multilingual emotion recognition

8. Unifying Conformal Language Tasks with In-Context Ensembles

Authors: Xiao Shi Huang, Chen-Yuan Lin, Bruce Kuwahara, Kin Kwan Leung, Jesse C. Cresswell arXiv: arxiv.org/abs/2609.03005 Summary: Many NLP tasks, such as summarization and extractive question answering, reduce to retrieving relevant content from documents under two constraints: coverage, retaining enough pertinent information to achieve some goal, and conciseness, removing as much irrelevant information as possible. Conformal prediction methods have been used to guarantee coverage, and must be optimized for conciseness through design of a score function. Trending because: 5 HuggingFace upvotes + combines conformal coverage guarantees with in-context ensembles

9. A Modular Agent for Reliable and Auditable Spatial Relation Verification in CT Scans

Authors: Simon Vincent Abel, Heiko Hillenhagen, Michael Götz, Timo Ropinski, Ayhan Can Erdur, Daniel Santak Wolf arXiv: arxiv.org/abs/2608.21140 Summary: Reliable spatial understanding is an important prerequisite for future medical vision-language systems that aim to support radiological report generation and structured image understanding. While modern vision-language models (VLMs) show promising performance on many medical imaging tasks, recent evidence suggests they remain weak in controlled spatial reasoning and often fail to reliably ground spatial relations in image evidence. Trending because: 5 HuggingFace upvotes + makes medical spatial reasoning modular, accurate, and auditable

10. LMSM: LLM Security Framework Inspired by Linux Security Modules

Authors: XiuYu Zhang, Bonan Ruan, Junfeng Fang, An Zhang, Tat-Seng Chua, Zhenkai Liang arXiv: arxiv.org/abs/2608.25697 Summary: Large language models (LLMs) are increasingly deployed with layered defenses, yet malicious prompts can still bypass them. Interpretability methods can expose model-internal signals along the generation path that could inform enforcement, but these signals are not security controls by themselves. Trending because: 5 HuggingFace upvotes + turns model-internal safety signals into a modular runtime enforcement framework

11. SpanCalib-VLM: Calibrated Hallucination Span Detection in Vision-Language Models

Authors: Amanuel Gizachew Abebe, Yasmin Moslem arXiv: arxiv.org/abs/2608.29974 Summary: Detecting hallucinations in Large Vision-Language Models (LVLMs) requires both accurate span localization and well-calibrated confidence scores. Fine-tuned generative VLMs excel at identifying hallucinated text spans but suffer from overconfidence and high inference latency. Trending because: 5 HuggingFace upvotes + targets both localization and confidence calibration for VLM hallucinations

12. ContextBias: Controlled Evaluation of Bias Persistence Under Context Shift in Text-to-Image Models

Authors: Shaghayegh Kolli, Sina Emami, Moreno D’Incà, Pouyan Nejadi, Nicu Sebe, Massimiliano Mancini, Jana Diesner arXiv: arxiv.org/abs/2608.29847 Summary: Text-to-image models learn associations between concepts - in the case of this paper, people’s professions, which we refer to as roles - and visual attributes. These associations can underpin many observed forms of stereotypical bias. Trending because: 5 HuggingFace upvotes + shows that text-to-image stereotypes persist under controlled context shifts

13. EvoGenUI-Bench: Evaluating LLMs as Multi-Turn Generative UI Assistants

Authors: Yue Peng, Lanke Xia, Zihan Wang, Jiahao Ye, Ke Ning, Hongyi Wen arXiv: arxiv.org/abs/2608.29387 Summary: Large language models can generate interactive web interfaces, but reliable generative UI requires maintaining an executable artifact as user requests evolve. We introduce EvoGenUI-Bench, a benchmark for multi-turn interface maintenance comprising 150 five-turn tasks and 750 turns across three scenarios: information presentation, executable interaction, and tool-grounded external state. Trending because: 5 HuggingFace upvotes + tests whether generated interfaces remain correct across multi-turn edits

14. RECAP-Forcing: Retaining Content Appearances for Long Video Generation

Authors: Haiyang Xu, Zheng Ding, Zhuowen Tu arXiv: arxiv.org/abs/2608.26671 Summary: Long autoregressive video generation faces a fundamental memory challenge: with a finite attention window, a model must decide which information from an ever-expanding history to retain. Existing methods organize memory temporally, preserving recent frames while compressing or discarding older ones. Trending because: 5 HuggingFace upvotes + uses appearance novelty rather than recency to preserve long-video consistency

15. DramaChain Bench: An End-to-End Benchmark for Short-Drama Generation

Authors: Haoyuan Shi, Mingtao Chen, Shuo Jiang, Ziyan Chen, Xuyi Sheng, Yiming Liu, Ying Zhang, Miao Wang, Jianxiang Lu, Fanyang Lu, Songyuanyi Lu, Xiele Wu, Zhichao Hu, Yuhong Liu, Richeng Xuan arXiv: arxiv.org/abs/2609.00646 Summary: Commercial short-drama production follows a multi-stage chain: script, storyboard, keyframe imagery, shot-level video, and the finished short drama. Most existing benchmarks evaluate solely the video-generation stage using pre-authored inputs instead of real upstream pipeline outputs. Trending because: 5 HuggingFace upvotes + evaluates the complete multi-stage short-drama generation pipeline

16. ReFlowSET: Representation-Aligned Latent Flow Matching for SAR-to-EO Image Translation

Authors: Jeonghyeok Do, Seungchul Lee, Munchurl Kim arXiv: arxiv.org/abs/2609.00968 Summary: SAR-to-EO image translation aims to generate electro-optical (EO) imagery from synthetic aperture radar (SAR) observations. Existing latent diffusion approaches typically inherit a predetermined autoencoder, although reconstruction fidelity can vary substantially across codecs and modalities. Trending because: 5 HuggingFace upvotes + aligns latent representations for higher-fidelity radar-to-optical translation

17. Token-Efficient Data Reasoning Agents via Adaptive Structuring of Unstructured Data

Authors: Milad Rezaei Hajidehi, Qitong Wang, Stratos Idreos arXiv: arxiv.org/abs/2608.31082 Summary: Valuable data remains embedded in unstructured sources: web pages, reports, contracts, filings, earnings calls, and PDFs. The big bet in enterprise AI is deploying LLM agents that reason over this data to answer complex questions for every knowledge worker. Trending because: 5 HuggingFace upvotes + reduces agent token costs by adaptively structuring reused evidence

18. Exploring Collaboration between a language and a non-language agent

Authors: Harini S I, Somesh Singh, Yaman K Singla, Rajiv Ratn Shah, David Doermann, Balaji Krishnamurthy arXiv: arxiv.org/abs/2609.00474 Summary: LLMs are increasingly deployed as orchestrators that coordinate specialized subagents to solve complex tasks through natural language. However, in many important domains like game playing and robotics, the strongest available agents are not language models. Trending because: 5 HuggingFace upvotes + studies direct latent collaboration between language and non-language agents

19. SnapBench: Benchmarking Snap-and-Ask Multimodal Retrieval for Mobile Interactions

Authors: Zirong Chen, Fuda Ye, Kuan Zhang, Enjun Du, Junfu Pu, Xinlei Wang, Xinyu Zuo, Lisheng Duan, Jin Ma, Yongqi Zhang arXiv: arxiv.org/abs/2608.29607 Summary: Mobile AI acts as a visual oracle, empowering users to snap a picture of something and ask for information. Snap-and-ask retrieval is now one of the most common entry points for mobile AI, yet photos are often blurry, while text questions may be short or mistyped. Trending because: 5 HuggingFace upvotes + benchmarks robust multimodal retrieval under realistic mobile input corruption

20. The Safeguard Worked. Is the LLM System Safer?

Authors: Pingyu Wu, Weiming Zhang, Nenghai Yu arXiv: arxiv.org/abs/2609.00519 Summary: Safeguards in deployed LLM services are evaluated by refusal, attack success, and policy violation rates. Those rates characterize how a control performed on the requests it was tested on. Trending because: 4 HuggingFace upvotes + distinguishes local safeguard scores from deployment-level safety evidence