The paper introduces SynOE-OD, a method for detecting out-of-distribution (OOD) objects in street scenes by generating synthetic outliers through text-guided inpainting and an open-vocabulary detector. These synthetic outliers are used to fine-tune a chosen object detector with an added OOD class, eliminating the need for auxiliary branches during inference. On the OoDIS benchmark, SynOE-OD achieves state‑of‑the‑art performance, outperforming the strongest open‑vocabulary baseline across multiple datasets while preserving competitive in‑distribution detection.
By Sadia Ilyas, Annika M\"utze, Klaus Friedrichs, Thomas Kurbiel, Matthias Rottmann
FlatClip is a method that leverages frozen image foundation models to predict fMRI activity by mapping cortical signals onto anatomy‑linked flatmaps and extracting features with a SigLIP2 encoder. Only a lightweight downstream probe is trained, avoiding fMRI‑specific encoder pretraining or fine‑tuning. Experiments on four resting‑state benchmarks show FlatClip surpasses ROI‑level baselines on HCP and ADNI, while its performance varies on PPMI and remains below the best voxel‑level models; on NSD visual‑fMRI decoding, restricting inputs to task‑active cortex improves recognition, and spatial perturbations confirm the importance of cortical arrangement and input‑region selection.
By Mo Wang, Wenhao Ye, Zihan Ning, Jiayu Zuo, Junfeng Xia, Hongkai Wen, Quanying Liu
The study investigates how well language models can generalize from fine‑tuned facts by distinguishing active facts—those that support further learning—from inert facts. Using a novel relabeling experiment (e.g., mapping Paris → Europe to Paris → crab), the authors fine‑tune models on a subset of subjects and test whether the learned mapping extends to unseen subjects sharing the same original answer. Across eleven models and 70 relations, they find a strong correlation (Pearson r = 0.87–0.92) between the linear readability of the original relation in pretrained representations and the model’s ability to generalize, a relationship that persists after controlling for various factors and remains robust to label randomization and semantic category changes.
By Faridoun Mehri, Hinrich Sch\"utze
The paper introduces On-Policy Reminder Distillation (OPRD), a method that enhances long-horizon coding agents by combining multi-scope reminder supervision with outcome-based routing. OPRD uses successful rollouts as autoregressive self-imitation targets and failed rollouts for reminder-conditioned reverse-KL supervision from a teacher that re-evaluates student prefixes. The approach improves task resolution rates by 11.3 points and cuts error events by 50.2% on several SWE-bench benchmarks.
By Yuechen Luo, Zhiyi Lai, Miao Ziqi, Zihang Xu, Jingxuan Xu, Songwei Yu, Wenqiang Zhu, Weihao Li, Junhao Fan, Haoyang Huang, Yanan Wu, Ken Deng
WEAVE is a framework that converts verified pull‑request histories into executable, dependency‑structured tasks for code agents. It organizes related PRs into bounded directed acyclic graphs (DAGs) using repository‑calibrated typed change loci, creating unified tasks that reflect real‑world code evolution. The framework also introduces Graph Credit, a deterministic, test‑grounded metric that measures partial progress consistent with dependencies, providing dense, verifiable rewards for reinforcement learning.
By Weihao Li, Zihang Xu, Zhongqiu Wang, Songwei Yu, Wenqiang Zhu, Yuechen Luo, Jingxuan Xu, Yanan Wu, Haoyang Huang, Ken Deng
EvolvingBrain introduces an adaptive skill system that enables multi-modal agents to self‑evolve in open worlds. The system comprises Stratified Skill Self‑Evolution, which extracts procedural and problematic patterns from trajectories to build a multi‑granularity skill architecture, and State‑Aware Adaptive Execution, which uses structured memory to guide skill fusion based on the agent’s current state. Experiments on VAB‑Minecraft and DiscoveryWorld show improved success rates and generalization, with code available for public release.
By Runxuan Liu, Jingqi Shi, Baoxin Wang, Honghong Zhao, Haoyuan Xu, Yueqing Sun, Jiafeng Liang, Tao He, Dayong Wu, Qi Gu, Shijin Wang, Ming Liu, Bing Qin
arXiv:2610. 21055v1 Announce Type: new Abstract: Recursive self-improvement, in which AI systems build stronger AI systems, is one promising route to superintelligence.
By Yufa Zhou, Yichen Wang, Jiakun Fan, Ziqi Shu, Zhichao Liu, Anru R. Zhang, Neiwen Ling, Zijun Wei
The paper studies how the distribution of information across chain‑of‑thought steps—measured by LP variance and Caveman compression ratio—relates to answer correctness in large language models. Across six model sizes, two training paradigms, and 60 GSM8K problems, it finds that lower LP variance predicts accuracy in supervised fine‑tuned models but not in reinforcement‑learning distilled ones, that models prefer more complex distributions, and that steering density at inference time harms accuracy while pre‑generation nudging and a new Entropy‑Gated Density Routing method improve it. The work demonstrates that natural density patterns are functionally important for reasoning and offers practical techniques to enhance performance.
By Tanush Kalhan, Anupam Chattopadhyay, Sriram Ranga, Amritendu Mukherjee
The paper presents a lightweight, scalable workflow for refining ontologies, demonstrated on the CASTON casting-domain ontology. It combines a fine‑tuned ontology embedding model to generate candidate axioms and replacement options with a Large Language Model (LLM) that proposes repair suggestions. In a small expert evaluation, the default setup—using a CASTON‑fine‑tuned embedding model and Qwen3‑8B—yielded high‑confidence repairs that were all accepted by domain experts.
By Hui Yang, Ziyu Li, Xuan Liu, Anders Jarfors, He Tan, Jiaoyan Chen
The paper introduces a tool‑augmented large language model framework that improves the reliability of medical calculations derived from free‑form patient notes. It separates clinical language understanding from rule execution by retrieving calculator specifications, generating executable tools, and using a constrained extractor to map notes to calculator inputs. On the MedCalc‑Bench dataset, the method achieves 86.73% final‑answer accuracy, outperforming one‑shot prompting and fine‑tuned baselines by 24.0 and 17.27 percentage points.
By Yang Zhong, Ling Luo, Xuanlyu Zhu, Yanhua Wang, Yuanyuan Sun, Jian Wang, Hongfei Lin
arXiv:2610. 21808v1 Announce Type: new Abstract: Weight quantization below four bits shrinks memory and bandwidth at a perplexity that looks acceptable, but often at the cost of factual knowledge and task accuracy.
By Johannes Knittel, Alec Helbling, Hanspeter Pfister
The paper introduces CART, a method for Multimodal Continual Instruction Tuning that coordinates adaptation between a projector and a language model by using task‑specific experts and weighted mixtures. CART jointly infers expert weights based on relevance and cross‑module compatibility, and it controls model growth through relevance‑guided expert consolidation. Experiments show that CART achieves state‑of‑the‑art performance on multimodal continual learning tasks.
By Ling-Xiang Zhang, Tao Hu, Da-Wei Zhou
MLExpert is a 35B Mixture-of-Experts agentic model designed to autonomously perform long‑horizon machine learning engineering (MLE) tasks. It uses a post‑training method that combines hierarchical agentic supervised fine‑tuning with Windowed Interaction RL to learn from intermediate outcomes in complex experimentation workflows. On the MLE‑Bench Lite benchmark, MLExpert raises the average medal rate from 37.9% to 72.7%, matching the performance of larger frontier models and demonstrating generalizable optimization capabilities across related tasks.
By Shangheng Du, Jinxin Shi, Shuaiyu Zhang, Yiming Wu, Junpeng Zhao, Shiyang Feng, Xin Li, Liang He, Xiangchao Yan, Bo Zhang
The paper introduces 3DCodeVerse, a research ecosystem that allows coding agents to generate editable and verifiable 3D content at scale. It comprises a large curated dataset of 3D executable code (3DCodeVerse‑2M), a multi‑agent harness that automatically evaluates and refines generated programs using execution checks and spatial measurements, and empirical studies that establish baselines and practical guidance for 3D coding. Experiments show that fine‑tuning on the dataset improves open‑source models’ 3D coding abilities, and the harness outperforms Claude Design in pairwise human evaluations when used with Claude Opus 5.5.
By Yipeng Gao, Ziyao Zeng, Jinfa Huang, Yijiang Li, Haolin Xiong, Wang Qin, Xinlei Yu, Youheng Yao, Jie Zhang, Kun-Yu Lin, Rong Liu, Yinyuan Zhao, Yajie Zhao, Meiqi Guo, Jiebo Luo, Lei Shu, Yunhao Ge, Xinyu Hu, Laurent Itti
LIBERO-PeRM is a benchmark for studying personalized robotic manipulation, focusing on how policies learn and follow user preferences in zero‑shot settings, from demonstrations, and across varying tasks and layouts. It introduces a taxonomy of language‑expressible preferences—motion, skill, and user‑profile choices—within the LIBERO simulation framework, generating 312 tasks with paired demonstrations and preference‑satisfaction metrics. Experiments reveal that current vision‑language‑action policies struggle with zero‑shot personalization, while supervised fine‑tuning improves learnability, yet challenges remain in generalization, decomposition, and conflict handling.
By Zhixu Li, Keqian Tang, Litian Gong, Jingyu Yao, Wenqian Zhang, Tianze Xu, Zehao Wang, Yuping Wang, Junge Zhang, Jiachen Li
EvoFlux is an inference‑time search method that repairs tool‑use plans generated by small language models (1.5B‑4B parameters). It iteratively edits structured graphs of tool calls, guided by execution errors, to replace tools, correct parameters, insert missing steps, or remove redundancies. On the MCP‑Bench dataset, EvoFlux raises execution success from about 3% to 17‑24% across four small planners, outperforming supervised fine‑tuning baselines and matching or exceeding zero‑shot performance.
By Kushal Raj Bhandari, Ling Yue, Ching-Yun Ko, Dhaval Patel, Shaowu Pan, Pin-Yu Chen, Jianxi Gao
The paper introduces Witness Overlap, a new method for determining directional provenance among open-weight model checkpoints. By using a third same-family checkpoint as a witness, the technique compares local geometry to infer which checkpoint is the parent of another, achieving 95.3% accuracy on 176 LLM checkpoints from 16 families. The authors also demonstrate robustness to weight noise and pruning, and extend the approach to VLM and diffusion families.
By Siyuan Li, Haoxuan Zeng, Xin Luo, Fernando Jia, Florence Li, Zhengyang Geng, Zico Kolter, Tai Sing Lee, Tianqin Li
UniFGVC introduces a training‑free framework for few‑shot fine‑grained visual classification that reframes the task as a multimodal retrieval problem. It uses a Category‑Discriminative Visual Captioner to generate attribute‑aware textual descriptions, then matches image‑description pairs with category templates via pre‑trained encoders. Experiments on 12 datasets show that UniFGVC consistently outperforms existing CLIP‑based and fully‑supervised MLLM methods, demonstrating the effectiveness of the retrieval paradigm.
By Haiyun Guo, Hongyu Guo, Xiangzhao Hao, Jiarui Guo, Jinqiao Wang, Tat-Seng Chua
The paper examines privacy risks in large language models (LLMs) through a unified threat model and systematic ablation study of four attacks: Membership Inference, Attribute Inference, Data Extraction, and Backdoor Attacks. It finds that membership inference is highly sensitive to dataset and attack strategy, attribute inference depends on model and training characteristics, data extraction grows with model size and repeated training data, and backdoor attacks consistently succeed. The authors provide practical recommendations for privacy‑aware LLM deployment, such as prioritizing membership leakage checks in retrieval‑augmented systems, reducing repeated personal data in training, monitoring attribute inference, and protecting fine‑tuning pipelines from backdoors.
By Karima Makhlouf, Lamiaa Basyoni, Syed Khaderi, Gabriel Marquez, Peter Sotomango, Mahmoud Awawdah, Sami Zhioua
The paper investigates why direct pairwise AUC surrogate training, though aligned with the ranking metric, can fail to provide corrective updates during stochastic training. It introduces the concept of Corrective Power, showing that in certain regions severe ranking errors persist without generating strong gradient signals, a phenomenon linked to degeneracies in the U‑statistics of pairwise gradients. The authors explain empirical observations—such as the effectiveness of squared‑hinge loss, the benefit of cross‑entropy warm‑up, and the advantage of parameter‑efficient fine‑tuning—through this theoretical lens, and validate their claims with experiments on eight computer‑vision and eight NLP tasks using various models.
By Jia Chen, Zhiyong Yang, Shilong Bao, Qianqian Xu, Qingming Huang