MERIT is a two‑stage framework for reviewer assignment that first trains a reviewer assessor using reinforcement learning to match paper‑specific expertise rubrics with reviewers’ prior work, guided by an LLM judge. The assessor’s predictions are then distilled into an embedding‑based retriever for efficient large‑scale assignment. Experiments show the 4B assessor outperforms larger general‑purpose LLMs on suitability classification, and the retriever achieves state‑of‑the‑art performance on LR‑Bench and the CMU Gold dataset.
By Zixuan Yang, Yibo Zhao, Weicong Liu, Xiang Li
The paper presents a pipeline that leverages large language models to extract grammatical rules, example sentences, and lexicons from descriptive grammar books, producing synthetic parallel corpora for fine‑tuning machine translation models. Evaluated on three low‑resource languages—Kalamang, Tuatschin, and Mandan—the synthetic data improves translation quality over seed‑data baselines in 75% of configurations for Kalamang and 59% for Tuatschin, achieving up to +8.8 ChrF++ gains. A factorial study across 96 configurations identifies which combinations of target part‑of‑speech, retrieval granularity, and sample volume drive performance gains and where they fail, demonstrating that static linguistic documentation can be repurposed for practical translation tools for severely under‑resourced languages.
By Varun Ghat Ravikumar, Sina Ahmadi, Lena J\"ager, Rico Sennrich
MisEdu‑RAG is a dual‑hypergraph retrieval‑augmented generation framework designed to help novice math teachers diagnose and remediate student misconceptions. It structures pedagogical knowledge as a concept hypergraph and real student mistake cases as an instance hypergraph, performing two‑stage retrieval to ground responses in both layers. On the MisstepMath dataset, MisEdu‑RAG outperforms baseline models, improving token‑F1 by 10.95% and achieving up to 15.3% higher quality across five dimensions, especially in diversity and empowerment.
By Zhihan Guo, Yuting Lu, Jionghao Lin
The paper introduces a densely labeled coastal dataset of over 1,000 images from Oahu, Hawaii, featuring 18 semantic classes and 7,400 annotated instances. Seven modern vision‑language models were evaluated using text‑to‑mask, mask‑to‑mask, and mask‑to‑text alignment experiments. Results show that broad landscape classes are recognized more accurately than conventional object and coastal classes, with coastal concepts posing the greatest challenge; however, performance differences are not solely due to environmental context, and alternative textual labels improve recognition of several coastal concepts.
By Seth Knoop, Chad R. Samuelson, Gabriel R. Slade, Brady Moon, Joshua G. Mangelson
CamPilot is a multi‑agent cinematic assistant that combines cinematographic planning with camera‑work control to generate more coherent and aesthetically pleasing movies from text prompts. It learns camera‑work planning from 14,000 professional films using a GRPO‑based learning paradigm, capturing motion patterns, composition principles, and cross‑shot relationships. The system is evaluated with a new benchmark, CamEval, and outperforms existing text‑to‑movie methods in cinematographic control and quality.
By Yang Wu, Stefano Petrangeli, Ishita Dasgupta, Yu Shen
The paper introduces Sewer-Transformer-ML, a hierarchical vision Transformer that fuses multi‑level features for multi‑label sewer defect classification, and two lightweight variants, Sewer-MobileNet-ML and Sewer-Mobile-TransNet, tailored for resource‑constrained inspection scenarios. On the Sewer‑ML test set, Sewer‑Transformer‑ML‑Base achieved an $F2_{ ext{CIW}}$ of 65.68% and an $F1_{ ext{Normal}}$ of 92.68%, topping the public leaderboard and surpassing the next best method by 7.6 percentage points in $F2_{ ext{CIW}}$. The lightweight Sewer‑MobileNet‑ML reached a comparable $F2_{ ext{CIW}}$ of 65.73% with only 17 M parameters, a 95% reduction from the base model, while Sewer‑Mobile‑TransNet achieved 96.43% accuracy under the standard data split, and ablation studies highlighted the effectiveness of direct concatenation for Transformer features and attention‑based fusion for multiscale CNN features.
By Xu Fang, Zhuoran Wang, Qing Li, Shengyu Zhang, Guanzhi Deng, Jianbiao He, Qingquan Li
LoopVAE introduces a recurrent depth architecture that reuses a scale‑ and loop‑conditioned core across different spatial scales while keeping resolution‑changing transitions separate. The four‑block core applies 28 block operations per encoder or decoder, enabling a 29M‑parameter convolutional model to achieve 0.28 rFID and 32.54 dB PSNR on ImageNet‑256 with roughly 65% fewer parameters than comparable VAEs. Experiments with both convolutional and Transformer operators, as well as ablations on parameter sharing, demonstrate competitive image quality metrics and reveal how targeted loop interventions and truncation affect reconstruction quality and computational trade‑offs.
By Zhiying Lu
V‑Retrver is an evidence‑driven retrieval framework that treats universal multimodal retrieval as an agentic reasoning process grounded in visual inspection. It allows multimodal large language models to selectively acquire visual evidence through external tools, alternating between hypothesis generation and targeted visual verification. The approach is trained with a curriculum that blends supervised activation, rejection‑based refinement, and reinforcement learning, achieving an average 23.0% improvement in retrieval accuracy across multiple benchmarks.
By Dongyang Chen, Chaoyang Wang, Dezhao Su, Xi Xiao, Zeyu Zhang, Jing Xiong, Qing Li, Yuzhang Shang, Shichao Kan
ABACUS is a 3‑billion‑parameter vision‑language model that simultaneously tackles object counting, crowd counting, referring‑expression counting, and count‑faithful image generation. It introduces density‑aware adaptive zooming with an objectness map, a boundary‑aware count policy trained via GRPO to avoid over‑ or under‑counting at crop edges, and a cycle‑consistent GRPO strategy that scores generated images for count accuracy and aesthetic quality without external critics. The model sets new state‑of‑the‑art performance on seven benchmarks, outperforming both specialized and larger generalist models.
By Anindya Mondal, Sauradip Nag, Anjan Dutta
The paper reports an entry to the EgoLongQA track of the Wearable‑AI Challenge at ECCV 2026 that achieved first place in the ≤2 B parameter division with a score of 0.8279 on the held‑out test set. The system is a single 2 B vision‑language model that answers multiple‑choice questions about ten‑minute egocentric videos in one greedy forward pass, obtained by distilling the junior perception module of a tool‑using agentic pipeline into a small student using teacher traces that answered correctly. By pruning the multilingual embedding table from 248,320 to 143,469 rows, the authors reduce the model to 1.9985 B parameters while maintaining identical logits on retained rows, enabling the model to reach 89 % of the accuracy of the large agentic pipeline with only 1.1 % of its parameters.
whyItMatters":"The work demonstrates that a highly efficient, sub‑2 B parameter vision‑language model can match the performance of a much larger agentic pipeline on long‑video perception tasks, showing the feasibility of deploying such models in resource‑constrained settings."
By Logesh Kumar Umapathi
Fortunate Recall (FR) introduces an ontology-driven policy layer that categorizes personal facts into over ten behavioral types and applies tailored lifecycle rules—such as differential decay, supersession, and event-time validity—to manage memory persistence in large language models. The FR-Bank implementation, independent of underlying infrastructure, achieves a 76.9% pass rate on the new LifecycleBench benchmark and improves LongMemEval-S performance, while significantly reducing confabulation rates compared to prior systems. Ablation studies show that the generic lifecycle metadata drives correctness, whereas the behavioral ontology enhances calibration and reduces downstream hallucinations.
By Ansuman Mullick, Eray T\"uz\"un
The paper investigates in-context binding errors in language models, showing that a linear probe can recover correct entity bindings from frozen hidden states even when the model outputs incorrect bindings. Across 16 checkpoints, the probe’s accuracy on failure cases surpasses a baseline by about 0.196, and a probe‑based score improves failure detection over the model’s confidence by 0.079 AUROC. Steering the residual stream toward the probe‑decoded binding further boosts accuracy by an average of 0.168 across eight models.
By Manas Venkata Sai Ravulapalli, Samrath Singh Chadha, Abhinav M. Hari
The paper revisits Kashin‑decomposition‑based weight quantization for large language models, introducing an improved algorithm that uses a sign‑randomized Discrete Cosine Transform (DCT) instead of a dense random orthogonal matrix. This change reduces per‑iteration cost from ≠(N^2) to ≠(N log N) and, combined with a greedy alternating‑update scheme, guarantees the four‑peak distribution needed for stable 2‑bit clustering while eliminating the need for multi‑restart k‑means. The resulting JAX pipeline, when paired with OPTQ‑style error compensation and QuIP‑style incoherence preprocessing, competes with state‑of‑the‑art quantization methods on OPT, Llama‑2, and Pythia at 4‑bit per channel, and remains numerically stable under stress configurations that cause other methods to diverge.
By Daria Cherniuk, Alexander Rudikov, Boris Kashin, Ivan Oseledets
The study analyzes 207,111 astronomy papers from 2015 to mid‑2026 to quantify how many contain language‑model‑generated vocabulary. Using a hierarchical Bayesian model calibrated on pre‑2020 unassisted papers and 392 papers that disclose model use, the authors estimate that in 2025 roughly 54% (±8% statistical, ±26% systematic) of papers show a language‑model trace, with the estimate remaining above 36% under various assumptions. Despite only 0.81% of 2025 papers explicitly declaring model assistance, the trace is pervasive, and the detectable signal is fading as authors adapt to the characteristic words.
whyItMatters":"The findings reveal that language‑model assistance has become widespread in recent astronomy research, yet most authors do not disclose its use, highlighting a growing gap between actual practice and transparency in scholarly writing."
By Serat M. Saad, Yuan-Sen Ting
Musec introduces MomentUm SpEctral Clipping, an optimizer-level, architecture‑agnostic technique that replaces Muon’s spectral flattening with selective spectral clipping to stabilize training. By clipping singular values above a threshold while preserving the momentum’s spectral structure, Musec addresses loss spikes and unbounded weight growth without requiring architecture‑specific changes. Soft Musec, an efficient implementation using smooth spectral saturation via coupled Newton‑Schulz iterations, offers convergence guarantees in nonconvex nonsmooth stochastic optimization and empirically improves stability across diverse learning rates and model sizes.
By Zhuanghua Liu, Menglian Wang, Luo Luo
The paper introduces a four‑layer architecture for Cognitive Digital Twins (CDTs) that extends traditional Digital Twin systems by embedding cognitive capabilities. It outlines a self‑evolving closed loop across physical, digital‑twin, cognitive, and task layers, enabling task‑specific cognitive models and decision generation that are refined through operational feedback. Two operation modes—user‑request‑driven and self‑driven cognition—are described, along with key enabling mechanisms and a lightweight simulation demonstrating closed‑loop feasibility and efficiency gains.
By Haoran Gao, An Li, Zhen Li, Jun Cai
AgentAudit is an open, extensible framework that evaluates the full lifecycle of AI agents, assessing planning, tool selection, execution, memory, and reasoning across ten dimensions such as instruction integrity, security, and alignment. Unlike existing benchmarks that focus on single aspects, AgentAudit analyzes the entire execution trace to attribute failures to specific stages. The framework was tested on five large language models, revealing significant differences in trustworthiness even among models with similar task‑completion performance.
By Shrey Nag, Sachita, Abhishek Kumar Singh, Lipi Goel, Rajeshwar Singh Janwar
Auto-RecSys is an autonomous research system designed to scale long-horizon experimentation for industry‑scale recommendation models. It tackles long feedback loops and system complexity by enabling distributed asynchronous execution, centralized cross‑server memory, and a cognitive‑procedural separation that combines natural‑language skill files with deterministic scripts. The system incorporates a dual‑loop self‑evolving architecture—Execution Evolution and Idea Evolution loops—to refine operational playbooks and guide future experiments, thereby reducing human effort per cycle and improving reliability as playbooks mature.
By Ming Li, Dai Li, Xuying Ning, Bo Sun, Rui Li, Yi Zhang, Silvia Gong, Xuan Cao, Rui Li, Cornelia Carapcea, Qunshu Zhang, Zhigang Wang, Yinglong Xia, Andy Wang
The paper investigates whether a language‑model agent can autonomously study an unfamiliar environment without prior task instructions or examples, and decide how to prepare for future tasks. It formalizes task‑agnostic environment preprocessing, where a studying system explores under a budget to produce reusable artifacts for a later solver. Experiments on six diverse benchmarks show that a meta‑agent variant often outperforms fixed methods, though larger budgets do not consistently boost downstream reward, yet the artifacts still reduce test‑time sampling needed to achieve a target score.
By Vinay Samuel, Varun Ursekar, Vijay S. Kalmath, Apaar Shanker, Veronica Chatrath, Yuan Xue
The paper introduces SurgicalRoomAgent, a voice‑interactive multi‑agent system for smart operating rooms that leverages large language models to understand natural language, control devices, record intraoperative events, and generate surgical reports. Its layered architecture includes a voice interaction pipeline (wake, ASR, turn detection, agent reasoning, TTS) and an agent core (skill registry, task planner, device manager). Three key technologies—KV Cache prefix warming, streaming partial JSON parsing with early parallel task execution, and progressive skill prompt disclosure—reduce latency and maximize context efficiency, enabling real‑time operation within a 16,384‑token limit.
By Tianxiang Zhou