Trending Papers

VibeVoice synthesizes long-form multi-speaker speech using next-token diffusion and a highly efficient continuous speech tokenizer, achieving superior performance and fidelity.

A large language model adapted for time-series forecasting achieves near-optimal zero-shot performance on diverse datasets across different time scales and granularities.

A multi-agent framework using large language models for stock trading simulates real-world trading firms, improving performance metrics like cumulative returns and Sharpe ratio.

A large-scale dynamic dataset derived from AAA games is introduced to improve generative inverse and forward rendering, featuring high-resolution synchronized RGB and G-buffer data alongside a novel VLM-based evaluation method that correlates well with human judgment.

PaddleOCR-VL, a vision-language model combining NaViT-style dynamic resolution and ERNIE, achieves state-of-the-art performance in document parsing and element recognition with high efficiency.

Enhancements to the AgentScope platform improve scalability, efficiency, and ease of use for large-scale multi-agent simulations through distributed mechanisms, flexible environments, and user-friendly tools.

AgentScope enhances agentic applications by providing flexible tool-based interactions, unified interfaces, and advanced infrastructure based on the ReAct paradigm, supporting efficient and safe development and deployment.

The AI Scientist-v2 autonomously proposes hypotheses, performs experiments, analyzes data, and writes peer-reviewed scientific papers, marking the first fully AI-generated paper accepted by a conference.

LightRAG improves Retrieval-Augmented Generation by integrating graph structures for enhanced contextual awareness and efficient information retrieval, achieving better accuracy and response times.

SKILL0 enables LLM agents to internalize skills during training, allowing zero-shot autonomous behavior through a dynamic curriculum that reduces contextual overhead while improving task performance.

Latent space is emerging as a fundamental computational substrate for language-based models, offering advantages over explicit token-level approaches through continuous representation that mitigates linguistic redundancy and sequential inefficiency.

LeWorldModel presents a stable end-to-end JEPA framework that trains efficiently from raw pixels using minimal loss terms while maintaining competitive performance in control tasks and encoding meaningful physical structures.

Agent Lightning is a flexible RL framework for training LLMs in various agents, using a hierarchical RL algorithm and decoupling execution from training to handle complex interactions.

PagedAttention algorithm and vLLM system enhance the throughput of large language models by efficiently managing memory and reducing waste in the key-value cache.

MinerU2.5, a 1.2B-parameter document parsing vision-language model, achieves state-of-the-art recognition accuracy with computational efficiency through a coarse-to-fine parsing strategy.

Mem0, a memory-centric architecture with graph-based memory, enhances long-term conversational coherence in LLMs by efficiently extracting, consolidating, and retrieving information, outperforming existing memory systems in terms of accuracy and computational efficiency.

A continual meta-learning framework for large language model agents that jointly evolves policies and reusable behavioral skills while minimizing downtime through opportunistic updates and skill-driven adaptation.

Hyperagents represent a self-referential framework that integrates task and meta-agents into a single editable program, enabling metacognitive self-modification and open-ended improvement across diverse computational domains.

OpenDevin is a platform for developing AI agents that interact with the world by writing code, using command lines, and browsing the web, with support for multiple agents and evaluation benchmarks.

AutoFigure-Edit is an end-to-end system that generates editable scientific illustrations from text descriptions and reference images, supporting flexible style adaptation and efficient refinement.

daVinci-MagiHuman is an open-source audio-video generative model that synchronizes text, video, and audio through a single-stream Transformer architecture, achieving high-quality human-centric content generation with efficient inference capabilities.

EvoScientist is an adaptive multi-agent framework that enhances scientific discovery by continuously learning from past interactions through persistent memory modules.

A centralized orchestrator dynamically directs LLM agents via reinforcement learning, achieving superior multi-agent collaboration in varying tasks with reduced computational costs.

RAG-Anything is a unified framework that enhances multimodal knowledge retrieval by integrating cross-modal relationships and semantic matching, outperforming existing methods on complex benchmarks.

QuantAgent, a multi-agent LLM framework, excels in high-frequency trading by leveraging specialized agents for technical indicators, chart patterns, trends, and risk, outperforming existing neural and rule-based systems.

Bitnet.cpp enhances edge inference for ternary LLMs using a novel mixed-precision matrix multiplication library, achieving significant speed improvements over baselines.

Efficient Universal Perception Encoder (EUPE) improves edge device performance by distilling knowledge from multiple vision encoders through a two-stage scaling approach, achieving superior representation quality compared to previous methods.

AutoDev is an AI-driven software development framework that automates complex engineering tasks within a secure Docker environment, achieving high performance in code and test generation.

SmolDocling is a compact vision-language model that performs end-to-end document conversion with robust performance across various document types using 256M parameters and a new markup format.

LlamaFactory is a unified framework enabling efficient fine-tuning of large language models across various tasks using a web-based user interface.

A novel GPT-based model, OmniFlatten, enables real-time natural full-duplex spoken dialogue through a multi-stage post-training technique that integrates speech and text without altering the original model's architecture.

OpenClaw-RL framework enables policy learning from diverse next-state signals across multiple interaction modalities using asynchronous training with PRM judges and hindsight-guided distillation.

DreamLite is a compact unified on-device diffusion model that supports both text-to-image generation and text-guided image editing with efficient training and inference.

A self-supervised framework optimizes prompts for both closed and open-ended tasks by evaluating LLM outputs without external references, reducing costs and required data.

UniDriveVLA is a unified vision-language-action model for autonomous driving that decouples spatial perception and semantic reasoning through a mixture-of-transformers architecture with expert coordination and progressive training.

Agentic coding tools receive goals written in natural language as input, break them down into specific tasks, and write or execute the actual code with minimal human intervention. Central to this process are agent context files ("READMEs for agents") that provide persistent, project-level instructions. In this paper, we conduct the first large-scale empirical study of 2,303 agent context files from 1,925 repositories to characterize their structure, maintenance, and content. We find that these files are not static documentation but complex, difficult-to-read artifacts that evolve like configuration code, maintained through frequent, small additions. Our content analysis of 16 instruction types shows that developers prioritize functional context, such as build and run commands (62.3%), implementation details (69.9%), and architecture (67.7%). We also identify a significant gap: non-functional requirements like security (14.5%) and performance (14.5%) are rarely specified. These findings indicate that while developers use context files to make agents functional, they provide few guardrails to ensure that agent-written code is secure or performant, highlighting the need for improved tooling and practices.

The Qwen3-TTS series presents advanced multilingual text-to-speech models with voice cloning and controllable speech generation capabilities, utilizing dual-track LM architecture and specialized speech tokenizers for efficient streaming synthesis.

Frontier large language models exhibit Internal Safety Collapse, where they generate harmful content under specific task conditions, revealing inherent vulnerabilities despite alignment efforts.

A unified conversational language model, CALM, enhances both multi-turn conversation management and API usage by leveraging a multi-task dataset, outperforming specialized models across various benchmarks.

DataFlex is a unified framework for dynamic data-centric training of large language models that supports sample selection, domain mixture adjustment, and sample reweighting while maintaining compatibility with standard training workflows and enabling efficient large-scale deployment.

ViGoR benchmark addresses limitations in current AIGC evaluation by introducing a comprehensive framework for assessing visual generative reasoning across multiple modalities and cognitive dimensions.

Large language models face limitations in adapting to changing real-world environments, necessitating a new approach called agentic evolution that treats deployment-time improvement as a goal-directed optimization process.

UniRecGen combines feed-forward reconstruction and diffusion-based generation in a shared canonical space to produce complete and consistent 3D models from sparse inputs through disentangled cooperative learning.

MemOS, a memory operating system for Large Language Models, addresses memory management challenges by unifying plaintext, activation-based, and parameter-level memories, enabling efficient storage, retrieval, and continual learning.

GEMS is an agent-native multimodal generation framework that enhances model capabilities through structured multi-agent optimization, persistent memory, and domain-specific skills across general and downstream tasks.

CARLA-Air integrates high-fidelity driving and multirotor flight simulation within a unified Unreal Engine framework, supporting joint air-ground agent modeling with photorealistic environments and multi-modal sensing capabilities.

LTX-2 is an open-source audiovisual diffusion model that generates synchronized video and audio content using a dual-stream transformer architecture with cross-modal attention and classifier-free guidance.

Memory Sparse Attention (MSA) enables large language models to process extremely long contexts with linear complexity and high efficiency through innovations like sparse attention and document-wise RoPE.

PackForcing enables efficient long-video generation through hierarchical KV-cache management and spatiotemporal compression while maintaining temporal consistency and reducing memory usage.

mmGRPO, a multi-module extension of GRPO, enhances accuracy in modular AI systems by optimizing LM calls and prompts across various tasks.