DEPICT is a new training‑free metric for evaluating text‑to‑image alignment. It replaces fixed reference answers with an agreement rule that compares image‑based and caption‑only responses, weighting questions by how decisively the caption determines them. By merging this agreement score with a holistic score, DEPICT improves negation accuracy dramatically and outperforms existing training‑free metrics while matching or exceeding fine‑tuned evaluators on several benchmarks.
By Vasco Ramos, Sandra Godinho Silva, Joao Magalhaes, Ricardo Rei, Pedro Henrique Martins
The paper introduces Adaptive Mutual Distillation (AMD), a post‑training framework that jointly trains two large language models using different task‑balancing strategies. AMD evaluates and selects distillation weight adjustments via short training probes and task‑wise validation, leading to models that outperform supervised fine‑tuning baselines on six benchmarks across three backbones. Merging the two AMD models further improves performance, surpassing multi‑task fine‑tuning by an average of 2.91 points.
By Baohang Li, Xiaocheng Feng, Yichong Huang, Chengpeng Fu, Wenshuai Huo, Zekun Zhou, Zekun Yuan, Tingjia Zhang, Bing Qin
MuLoRA introduces a spectrally balanced low‑rank adaptation method for continual learning, addressing the issue of spectral plasticity collapse where adaptation energy concentrates in few singular modes. It uses historical whitening to identify task‑adaptive input directions and approximate polar orthogonalization of momentum updates to reduce spectral concentration, ensuring effective utilization of the low‑rank space. Experiments on five class‑incremental benchmarks and eight incremental settings show MuLoRA achieving the highest mean accuracy in 15 of 16 metrics.
By Junkang Liu
The paper investigates how to reduce the per-user adaptation burden in personalized large language models by separating reusable personalization capacity from user-specific adjustments. Through empirical studies, it shows that shared low‑rank factors can capture much of the cross‑user structure, while a tiny user code suffices for individual correction. The proposed LINEUP framework achieves state‑of‑the‑art performance on six personalized tasks while using only eight scalars per user compared to millions of private‑LoRA parameters.
By Songyuan Sui, Srikanth Malla, Chiho Choi, Joon Hee Choi
This paper introduces a personalized automatic speech recognition system for a Czech speaker with a permanent tracheal stoma and severe dysarthria. The authors release a 33‑hour annotated dataset collected via an artificial conversation protocol and develop a multi‑stage training pipeline based on Whisper Base, fine‑tuning on Czech speech, simulated tracheostomic speech, and the speaker’s data. Evaluations in scripted, question‑answering, and spontaneous dialogue scenarios show a 50 % relative reduction in character error rate compared to the Whisper Base baseline and better accuracy than the speaker’s assistants on isolated utterances.
By David Nadrchal, Monorama Swain, Florian Schmid, Gerhard Widmer, Paul Primus
The Writerslogic team participated in the CLEF 2026 SimpleText shared task, tackling both text simplification (Task 1) and complexity spotting (Task 2). For simplification, they built a multi‑candidate pipeline with GPT‑4o‑mini, selecting the best candidate via a reference‑free heuristic, and their Claude Sonnet 4 submission achieved a SARI of 47.43 and BLEU of 14.21, ranking third overall on the Task 1 leaderboard. For complexity spotting, they fine‑tuned a DeBERTa‑v3‑large NLI model on 350 K labeled pairs, achieving a macro F1 of 0.8081 (0.8085 in an ensemble) on binary over‑generation identification and 0.804 accuracy on multi‑class error classification, placing them second among unique teams.
By David L. Condrey
The paper introduces Percept-V, a dataset of 6,000 program-generated images across 30 domains designed to test simple visual perception skills from the TVPS-4 framework. Experiments show that state‑of‑the‑art multimodal large language models perform poorly compared to humans, especially as image complexity increases, and that fine‑tuning yields only limited generalization to related datasets. The study highlights specific perception skills that remain challenging for current models.
By Samrajnee Ghosh, Ashish Goswami, Naman Agarwal, Hemanshu Garg, Chinmay Mittal, Mausam, Parag Singla
The paper introduces MITE, a method that transforms biomedical named entity recognition (BioNER) into a structure‑to‑structure generation task by encoding instructions and outputs in multiple programming languages (Python, C++, Java). This approach provides structurally diverse supervision without extra biomedical knowledge, and during inference it aggregates predictions via entity‑level voting to reduce language‑specific variance. Experiments on six BioNER datasets show that MITE outperforms BERT‑based and LLM‑based baselines and generalizes well across datasets.
By Songtao Li, Yijia Zhang, Jianyuan Yuan, Shidi Zhang, Fengyu Zhang, Hongfei Lin
The paper introduces a two‑stage training framework for rubric‑based reinforcement learning (RL) that first uses rubric‑privileged on‑policy distillation (RP‑OPD) to provide dense token‑level supervision, then applies RL to optimize the rubric reward. In RP‑OPD, a student model learns to match a rubric‑aware teacher’s next‑token distributions on student‑generated prefixes. Experiments on HealthBench, ResearchQA, and RubricHub Science show that this two‑stage approach yields the highest scores and reduces reward hacking compared to a standard SFT + RL baseline.
By Xinpeng Wang, Wei Shi, Yu-Chia Chen, Maria Zontak, Yun He, Richard Yuanzhe Pang
The paper introduces TEGER, a residual covariance model that updates a forecaster’s joint predictive uncertainty at test time using closed‑form Gaussian conditioning on recent residuals and an exponential moving‑average volatility term. It leverages a fixed sensor graph to encode spatial correlations without retraining the underlying forecasting backbone. Experiments on four traffic datasets show significant improvements in 60‑minute CRPS, and applying TEGER to a frozen foundation model (Chronos) reduces CRPS from 0.1798 to 0.1736.
By Seyed Mohamad Moghadas, Esther Rodrigo Bonet, Bruno Cornelis, Adrian Munteanu
StanceEval 2026 is the second edition of a shared task on stance detection in Arabic social media, where systems must classify a tweet’s stance toward a target as Favor, Against, or None. The event featured two tracks: Track 1 tests cross‑target transfer on thematically related topics (Women Driving vs. Women Empowerment), while Track 2 evaluates cross‑domain transfer to entirely unseen targets (E‑Cars and Trimester System). With 80 registered teams and 30 submissions, top systems achieved $F_{avg2}$ scores of 0.8994 (Track 1) and 0.9400 (Track 2), surpassing baseline performance and highlighting challenges such as target polarization and dialectal nuance.
By Rasha Albalawi, Nuha Albadi, Hamzah Luqman, Asma Yamani, Maram Kurdi, Saad Ezzini, Ahmed Ashraf, Maged Al-Shaibani, Nora Alturayeif
SCOPE-4D is an endoscopic 4D geometry foundation model that predicts camera parameters, dense geometry, and 3D tissue trajectories from monocular RGB video in a single forward pass. The authors introduce SCOPE-5K, a curated dataset of about 5,000 real and synthetic gastrointestinal endoscopy and laparoscopy clips, and use it for geometric supervised fine‑tuning. Adding Common–Residual Motion (CRM) constraints and trajectory supervision further improves camera and depth estimation and enables dense 3D tissue tracking, as shown by evaluations on public and new benchmarks and a blinded user study.
By Chaoyi Zhou, Zhongpai Gao, Anwesa Choudhuri, Meng Zheng, Benjamin Planche, Run Wang, Terrence Chen, Siyu Huang, Ziyan Wu
The paper introduces Behavior Pack Optimization (BPO), a post‑training method for video multimodal large language models that replaces single‑response rewards with a set of outputs across counterfactual views. BPO enforces stability when interventions are irrelevant, sensitivity when key evidence is removed, and abstention when no evidence remains, using an anchor‑relative advantage to keep the objective stable with small pack sizes. Experiments on datasets such as TempCompass, MVBench, and NExT‑QA show that BPO improves macro accuracy and abstention metrics for models like Qwen2.5‑VL‑7B‑Instruct, with gains that transfer to other benchmarks and models.
By Zhaolu Kang, Shiyu Liu, Tailong Luo, Wei Zhang, Yingjie He, Lei Wei, Guansu Wang, Liang He, Siheng Wang, Guangyuan Dong, Jiaqi Su, Shuang Chen, Haoyu Ji, Qishi Zhan, Kaiyue Zhou
The paper introduces SalTempto, a new video saliency benchmark featuring 224 highly dynamic clips from HACS‑Segments, each with event context and gaze data from up to 16 subjects. It shows that static baselines perform poorly on SalTempto compared to LEDOV, while fine‑tuned temporal models achieve significant gains, yet still leave a large portion of explainable gaze information unaccounted for. The dataset highlights the need for better temporal modeling in video saliency research.
By Susmit Agrawal, Rebecca Wanner, Juliane Verwiebe, Matthias Tangemann, Matthias Bethge, Matthias K\"ummerer
The paper investigates how affective fine‑tuning shapes the internal architecture of multimodal foundation models. By analyzing 13 model instances across nine designs, it finds that adapting the feed‑forward network (FFN) consistently outperforms attention‑only adaptation and nearly matches full‑model tuning, revealing the FFN as an efficient adaptation substrate. Moreover, joint optimization leads to emergent functional specialization, notably a prominent gate projection pathway, which the authors exploit in Gate‑Focused Efficient Tuning (GET) to achieve 96.2–98.0% of full‑model performance with only 19.3–24.5% of the trainable parameters.
By Zhen Zhang, Runhao Zeng, Sicheng Zhao, Xiping Hu
The paper addresses the challenge of selecting external datasets for private transfer learning by modeling high‑dimensional regression with heterogeneous sources and a weighted ridge estimator. It relies solely on aggregated statistics and offers privacy guarantees under $
ho$‑zero‑concentrated differential privacy for labels or both features and labels. A deterministic equivalent of test error is derived, enabling optimization of hyperparameters and decision‑making about the utility of private external data without accessing individual records.
By Filip Kova\v{c}evi\'c, Edwige Cyffers, Stefano Sarao Mannelli, Marco Mondelli
arXiv:2610.03419v1 Announce Type: new
Abstract: In this work we formulate ultrasound multistatic recovery from arbitrary transmit sequences as a Bayesian inference problem. To that end, we train a de...
By Simon Penninga, Ruud van Sloun
arXiv:2610.03503v1 Announce Type: new
Abstract: Training-free posterior sampling methods, also known as Plug-and-Play methods, leverage pretrained unconditional diffusion or flow-matching models to s...
By Liam Moroy, Jean-Fran\c{c}ois Giovannelli, Yoann Altmann, Steve McLaughlin, Fr\'ed\'eric Champagnat, Guillaume Bourmaud
arXiv:2610.03641v1 Announce Type: new
Abstract: Masked discrete diffusion models offer a promising alternative to autoregressive generation, but iterative sampling can be costly, and intractable sequ...
By Vladislav Gromadskii, David Li, Samson Gourevitch, Yazid Janati, Eric Moulines, Maxim Panov, Alexander Korotin
arXiv:2610.03702v1 Announce Type: new
Abstract: The choice of post-training data for large language models substantially affects downstream performance. Gradient-based data selection is a popular app...
By Lyuxin David Zhang, Eric Wong, Surbhi Goel, Anton Xue