The paper systematically evaluates concept-based explainability in MLLMs, finding that forcing models to generate formal explanations degrades predictive accuracy, suggesting that explaining is genuin…

This paper analyzes the poor performance of Meta-learning for Training-data Selection (MTS) and proposes that increasing the batch size and incorporating informative features can significantly improve…

CORE-MTL proposes a representation-centric framework that uses causal orthogonal representations to disentangle task-relevant structure from nuisance variation in multi-task learning, achieving superi…

The paper proposes AlignG, a method that learns context-conditioned predicate semantics by using prototype feedback to adapt relation representations based on image-specific evidence, significantly im…

The paper investigates the limits of self-evolution in LLM reasoning under closed-loop settings, finding that while self-improvement is significant, it consistently falls short of perfect oracle super…

ExpWeaver introduces a novel framework for LLM agents to learn from past experiences using latent retrieval-augmented generation, achieving state-of-the-art performance while significantly improving t…

GiPL proposes a novel two-branch framework combining iterative pseudo-label self-training and generative data augmentation to significantly improve Cross-Domain Few-Shot Object Detection by better uti…

The paper proposes Visual Gradient Steering (VGS), a method that decomposes the distillation loss into language and visual components and steers the optimization to prioritize visual grounding, signif…

The paper introduces a novel two-stage framework to achieve robust, category-agnostic object localization in-context (ICL) by optimizing attention and minimizing localization error using reinforcement…

ROVER is a lightweight, learnable plugin that efficiently routes and integrates object-centric visual evidence across multiple images and objects, significantly improving performance on grounded multi…

The paper introduces Multi-Response Training (MRT) to combat the 'mode lottery' problem in language model fine-tuning, showing that retaining multiple valid responses significantly improves distributi…

The paper introduces BabyCL, a continual multimodal learning framework that processes egocentric video data in a single chronological pass, demonstrating that meaningful word-referent mappings can be…

The paper introduces OmniVerifier-M1, a multimodal meta-verifier that uses symbolic outputs and decoupled reinforcement learning to provide robust, fine-grained verification and error localization for…

The paper introduces Contrastive Reflection (CORE), a novel non-parametric method that rapidly improves language model reasoning by distilling contrasts between successful and unsuccessful problem att…

This paper systematically investigates unlearnable examples (UEs) across diverse training paradigms, finding that existing UEs fail under pretraining-finetuning (PF) settings, and proposes Shallow Sem…

The paper introduces OCC-RAG, a family of compact, task-specialized Small Language Models (SLMs) designed to achieve highly faithful, multi-hop question answering grounded strictly in provided context…

BayesNCL introduces a probabilistic gating mechanism to resolve the optimization conflict in Contrastive Learning, leading to highly disentangled and semantically consistent representations.