Graph-GRPO: Stabilizing Multi-Agent Topology Learning via Group Relative Policy Optimization
arXiv:2603.02701v1 Announce Type: new Abstract: Optimizing communication topology is fundamental to the efficiency and effectiveness of Large Language Model (LLM)-based Multi-Agent Systems (MAS). While recent approaches utilize reinforcement learning to dynamically construct task-specific graphs, they typically rely on single-sample policy...
OCR or Not? Rethinking Document Information Extraction in the MLLMs Era with Real-World Large-Scale Datasets
arXiv:2603.02789v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) enhance the potential of natural language processing. However, their actual impact on document information extraction remains unclear. In particular, it is unclear whether an MLLM-only pipeline--while simpler--can truly match the...
Faster, Cheaper, More Accurate: Specialised Knowledge Tracing Models Outperform LLMs
arXiv:2603.02830v1 Announce Type: new Abstract: Predicting future student responses to questions is particularly valuable for educational learning platforms where it enables effective interventions. One of the key approaches to do this has been through the use of knowledge tracing (KT)...
A Browser-based Open Source Assistant for Multimodal Content Verification
arXiv:2603.02842v1 Announce Type: new Abstract: Disinformation and false content produced by generative AI pose a significant challenge for journalists and fact-checkers who must rapidly verify digital media information. While there is an abundance of NLP models for detecting credibility signals...
Learning to Generate and Extract: A Multi-Agent Collaboration Framework For Zero-shot Document-level Event Arguments Extraction
arXiv:2603.02909v1 Announce Type: new Abstract: Document-level event argument extraction (DEAE) is essential for knowledge acquisition, aiming to extract participants of events from documents.In the zero-shot setting, existing methods employ LLMs to generate synthetic data to address the challenge posed by...
MaBERT:A Padding Safe Interleaved Transformer Mamba Hybrid Encoder for Efficient Extended Context Masked Language Modeling
arXiv:2603.03001v1 Announce Type: new Abstract: Self attention encoders such as Bidirectional Encoder Representations from Transformers(BERT) scale quadratically with sequence length, making long context modeling expensive. Linear time state space models, such as Mamba, are efficient; however, they show limitations in...
PrivMedChat: End-to-End Differentially Private RLHF for Medical Dialogue Systems
arXiv:2603.03054v1 Announce Type: new Abstract: Large language models are increasingly used for patient-facing medical assistance and clinical decision support, but adapting them to clinical dialogue often requires supervision derived from doctor-patient conversations that may contain sensitive information. Conventional supervised fine-tuning...
Compact Prompting in Instruction-tuned LLMs for Joint Argumentative Component Detection
arXiv:2603.03095v1 Announce Type: new Abstract: Argumentative component detection (ACD) is a core subtask of Argument(ation) Mining (AM) and one of its most challenging aspects, as it requires jointly delimiting argumentative spans and classifying them into components such as claims and...
Evaluating Performance Drift from Model Switching in Multi-Turn LLM Systems
arXiv:2603.03111v1 Announce Type: new Abstract: Deployed multi-turn LLM systems routinely switch models mid-interaction due to upgrades, cross-provider routing, and fallbacks. Such handoffs create a context mismatch: the model generating later turns must condition on a dialogue prefix authored by a...
APRES: An Agentic Paper Revision and Evaluation System
arXiv:2603.03142v1 Announce Type: new Abstract: Scientific discoveries must be communicated clearly to realize their full potential. Without effective communication, even the most groundbreaking findings risk being overlooked or misunderstood. The primary way scientists communicate their work and receive feedback from...
BeyondSWE: Can Current Code Agent Survive Beyond Single-Repo Bug Fixing?
arXiv:2603.03194v1 Announce Type: new Abstract: Current benchmarks for code agents primarily assess narrow, repository-specific fixes, overlooking critical real-world challenges such as cross-repository reasoning, domain-specialized problem solving, dependency-driven migration, and full-repository generation. To address this gap, we introduce BeyondSWE, a comprehensive...
Learning When to Act or Refuse: Guarding Agentic Reasoning Models for Safe Multi-Step Tool Use
arXiv:2603.03205v1 Announce Type: new Abstract: Agentic language models operate in a fundamentally different safety regime than chat models: they must plan, call tools, and execute long-horizon actions where a single misstep, such as accessing files or entering credentials, can cause...
Using Learning Progressions to Guide AI Feedback for Science Learning
arXiv:2603.03249v1 Announce Type: new Abstract: Generative artificial intelligence (AI) offers scalable support for formative feedback, yet most AI-generated feedback relies on task-specific rubrics authored by domain experts. While effective, rubric authoring is time-consuming and limits scalability across instructional contexts. Learning...
Self-Play Only Evolves When Self-Synthetic Pipeline Ensures Learnable Information Gain
arXiv:2603.02218v1 Announce Type: cross Abstract: Large language models (LLMs) make it plausible to build systems that improve through self-evolving loops, but many existing proposals are better understood as self-play and often plateau quickly. A central failure mode is that the...
Routing Absorption in Sparse Attention: Why Random Gates Are Hard to Beat
arXiv:2603.02227v1 Announce Type: cross Abstract: Can a transformer learn which attention entries matter during training? In principle, yes: attention distributions are highly concentrated, and a small gate network can identify the important entries post-hoc with near-perfect accuracy. In practice, barely....
Safety Training Persists Through Helpfulness Optimization in LLM Agents
arXiv:2603.02229v1 Announce Type: cross Abstract: Safety post-training has been studied extensively in single-step "chat" settings where safety typically refers to refusing harmful requests. We study an "agentic" (i.e., multi-step, tool-use) setting where safety refers to harmful actions directly taken by...
HELIOS: Harmonizing Early Fusion, Late Fusion, and LLM Reasoning for Multi-Granular Table-Text Retrieval
arXiv:2603.02248v1 Announce Type: cross Abstract: Table-text retrieval aims to retrieve relevant tables and text to support open-domain question answering. Existing studies use either early or late fusion, but face limitations. Early fusion pre-aligns a table row with its associated passages,...
MUSE: A Run-Centric Platform for Multimodal Unified Safety Evaluation of Large Language Models
arXiv:2603.02482v1 Announce Type: cross Abstract: Safety evaluation and red-teaming of large language models remain predominantly text-centric, and existing frameworks lack the infrastructure to systematically test whether alignment generalizes to audio, image, and video inputs. We present MUSE (Multimodal Unified Safety...
FlashEvaluator: Expanding Search Space with Parallel Evaluation
arXiv:2603.02565v1 Announce Type: cross Abstract: The Generator-Evaluator (G-E) framework, i.e., evaluating K sequences from a generator and selecting the top-ranked one according to evaluator scores, is a foundational paradigm in tasks such as Recommender Systems (RecSys) and Natural Language Processing...
StitchCUDA: An Automated Multi-Agents End-to-End GPU Programing Framework with Rubric-based Agentic Reinforcement Learning
arXiv:2603.02637v1 Announce Type: cross Abstract: Modern machine learning (ML) workloads increasingly rely on GPUs, yet achieving high end-to-end performance remains challenging due to dependencies on both GPU kernel efficiency and host-side settings. Although LLM-based methods show promise on automated GPU...
RxnNano:Training Compact LLMs for Chemical Reaction and Retrosynthesis Prediction via Hierarchical Curriculum Learning
arXiv:2603.02215v1 Announce Type: new Abstract: Chemical reaction prediction is pivotal for accelerating drug discovery and synthesis planning. Despite advances in data-driven models, current approaches are hindered by an overemphasis on parameter and dataset scaling. Some methods coupled with evaluation techniques...
Subspace Geometry Governs Catastrophic Forgetting in Low-Rank Adaptation
arXiv:2603.02224v1 Announce Type: new Abstract: Low-Rank Adaptation (LoRA) has emerged as a parameter-efficient approach for adapting large pre-trained models, yet its behavior under continual learning remains poorly understood. We present a geometric theory characterizing catastrophic forgetting in LoRA through the...
Scaling Reward Modeling without Human Supervision
arXiv:2603.02225v1 Announce Type: new Abstract: Learning from feedback is an instrumental process for advancing the capabilities and safety of frontier models, yet its effectiveness is often constrained by cost and scalability. We present a pilot study that explores scaling reward...
Generalized Discrete Diffusion with Self-Correction
arXiv:2603.02230v1 Announce Type: new Abstract: Self-correction is an effective technique for maintaining parallel sampling in discrete diffusion models with minimal performance degradation. Prior work has explored self-correction at inference time or during post-training; however, such approaches often suffer from limited...
Adaptive Personalized Federated Learning via Multi-task Averaging of Kernel Mean Embeddings
arXiv:2603.02233v1 Announce Type: new Abstract: Personalized Federated Learning (PFL) enables a collection of agents to collaboratively learn individual models without sharing raw data. We propose a new PFL approach in which each agent optimizes a weighted combination of all agents'...
Length Generalization Bounds for Transformers
arXiv:2603.02238v1 Announce Type: new Abstract: Length generalization is a key property of a learning algorithm that enables it to make correct predictions on inputs of any length, given finite training data. To provide such a guarantee, one needs to be...
Boosting Meta-Learning for Few-Shot Text Classification via Label-guided Distance Scaling
arXiv:2603.02267v1 Announce Type: new Abstract: Few-shot text classification aims to recognize unseen classes with limited labeled text samples. Existing approaches focus on boosting meta-learners by developing complex algorithms in the training stage. However, the labeled samples are randomly selected during...
A Comparative Study of UMAP and Other Dimensionality Reduction Methods
arXiv:2603.02275v1 Announce Type: new Abstract: Uniform Manifold Approximation and Projection (UMAP) is a widely used manifold learning technique for dimensionality reduction. This paper studies UMAP, supervised UMAP, and several competing dimensionality reduction methods, including Principal Component Analysis (PCA), Kernel PCA,...
Temporal Imbalance of Positive and Negative Supervision in Class-Incremental Learning
arXiv:2603.02280v1 Announce Type: new Abstract: With the widespread adoption of deep learning in visual tasks, Class-Incremental Learning (CIL) has become an important paradigm for handling dynamically evolving data distributions. However, CIL faces the core challenge of catastrophic forgetting, often manifested...
Preconditioned Score and Flow Matching
arXiv:2603.02337v1 Announce Type: new Abstract: Flow matching and score-based diffusion train vector fields under intermediate distributions $p_t$, whose geometry can strongly affect their optimization. We show that the covariance $\Sigma_t$ of $p_t$ governs optimization bias: when $\Sigma_t$ is ill-conditioned, and...