6 citations · 7 across the 24 of their papers we have counts for
27 papers
What Matters in On-Policy Distillation? A Perspective on Data Efficiency and Data Selection
Zhinan Hou, Jiaqi Zhang, Xunliang Cai +1
On-Policy Distillation (OPD) has emerged as a widely adopted post-training paradigm for enhancing large language models in reasoning domains. However, the data-centric mechanisms i…
LongCat Sparse Attention: Taming the Lightning via Streaming-aware Hierarchical Cross-Layer Indexing
Wen Zan, Jiaqi Zhang, Jianchao Tan +11
DeepSeek Sparse Attention (DSA) enables efficient long-context modeling through its Lightning Indexer. However, practical deployment remains constrained by the indexer's expensive…
Dyna-DINO: Efficient ViT Distillation Via Adaptive Representation Anchoring
Jiaqi Zhang, Ashton Lee, Anthony Wong +3
Vision Foundation Models (VFMs) with Vision Transformer (ViT) backbones, such as DINOv2, have become essential for downstream tasks like object recognition and semantic segmentatio…
ATLAS: All-round Testing of Long-context Abilities across Scales
Deli Huang, Cunguang Wang, Hongyin Tang +15
Long-context language models now advertise context windows up to millions of tokens, yet evaluations typically report a single length or a narrow task family, masking two failure m…
MONA: Muon Optimizer with Nesterov Acceleration for Scalable Language Model Training
Jiacheng Li, Jianchao Tan, Hongtao Xu +5
The Muon optimizer has recently offered a promising alternative to AdamW for large language model training, leveraging matrix orthogonalization to produce geometry-aware updates. H…
Iterative Multimodal Retrieval-Augmented Generation for Medical Question Answering
Xupeng Chen, Binbin Shi, Chenqian Le +5
Medical retrieval-augmented generation (RAG) systems typically operate on text chunks extracted from biomedical literature, discarding the rich visual content (tables, figures, str…
AsyncTLS: Efficient Generative LLM Inference with Asynchronous Two-level Sparse Attention
Yuxuan Hu, Jianchao Tan, Jiaqi Zhang +7
Long-context inference in LLMs faces the dual challenges of quadratic attention complexity and prohibitive KV cache memory. While token-level sparse attention offers superior accur…
FG-GDN: Enhancing Long-Context Gated Delta Networks with Doubly Fine-Grained Control
Pingwei Sun, Yuxuan Hu, Jianchao Tan +6
Linear attention mechanisms have emerged as promising alternatives to softmax attention, offering linear-time complexity during inference. Recent advances such as Gated DeltaNet (G…
LongCat-Next: Lexicalizing Modalities as Discrete Tokens
Meituan LongCat Team, Bin Xiao, Chao Wang +86
The prevailing Next-Token Prediction (NTP) paradigm has driven the success of large language models through discrete autoregressive modeling. However, contemporary multimodal syste…
Fast Catch-Up, Late Switching: Optimal Batch Size Scheduling via Functional Scaling Laws
Jinbo Wang, Binghui Li, Zhanpeng Zhou +5
Batch size scheduling (BSS) plays a critical role in large-scale deep learning training, influencing both optimization dynamics and computational efficiency. Yet, its theoretical f…
Scaling Embeddings Outperforms Scaling Experts in Language Models
Hong Liu, Jiaqi Zhang, Chao Wang +13
While Mixture-of-Experts (MoE) architectures have become the standard for sparsity scaling in large language models, they increasingly face diminishing returns and system-level bot…
QARM V2: Quantitative Alignment Multi-Modal Recommendation for Reasoning User Sequence Modeling
Tian Xia, Jiaqi Zhang, Yueyang Liu +25
With the evolution of large language models (LLMs), there is growing interest in leveraging their rich semantic understanding to enhance industrial recommendation systems (RecSys).…
OneMall: One Architecture, More Scenarios -- End-to-End Generative Recommender Family at Kuaishou E-Commerce
Kun Zhang, Jingming Zhang, Wei Cheng +29
In the wave of generative recommendation, we present OneMall, an end-to-end generative recommendation framework tailored for e-commerce services at Kuaishou. Our OneMall systematic…
The Law of Multi-Model Collaboration: Scaling Limits of Model Ensembling for Large Language Models
Dakuan Lu, Jiaqi Zhang, Cheng Yuan +2
Recent advances in large language models (LLMs) have been largely driven by scaling laws for individual models, which predict performance improvements as model parameters and data…
SpanNorm: Reconciling Training Stability and Performance in Deep Transformers
Chao Wang, Bei Li, Jiaqi Zhang +8
The success of Large Language Models (LLMs) hinges on the stable training of deep Transformer architectures. A critical design choice is the placement of normalization layers, lead…
In-Situ Tweedie Discrete Diffusion Models
Xiao Li, Jiaqi Zhang, Shuxiang Zhang +3
While diffusion models excel at generating continuous data such as images, adapting them to discrete tasks has relied on indirect approaches that either operate in continuous embed…
Optimizing Native Sparse Attention with Latent Attention and Local Global Alternating Strategies
Yuxuan Hu, Jianchao Tan, Jiaqi Zhang +7
In this work, we conduct a systematic analysis of Native Sparse Attention (NSA) and propose targeted improvements that enhance long-context modeling. A key insight is that alternat…
LongCat-Flash Technical Report
Meituan LongCat Team, Bayan, Bei Li +179
We introduce LongCat-Flash, a 560-billion-parameter Mixture-of-Experts (MoE) language model designed for both computational efficiency and advanced agentic capabilities. Stemming f…
Introducing LongCat-Flash-Thinking: A Technical Report
Meituan LongCat Team, Anchun Gui, Bei Li +122
We present LongCat-Flash-Thinking, an efficient 560-billion-parameter open-source Mixture-of-Experts (MoE) reasoning model. Its advanced capabilities are cultivated through a metic…
FastDINOv2: Frequency Based Curriculum Learning Improves Robustness and Training Speed
Jiaqi Zhang, Juntuo Wang, Zhixin Sun +2
Large-scale vision foundation models such as DINOv2 boast impressive performances by leveraging massive architectures and training datasets. But numerous scenarios require practiti…
GradPower: Powering Gradients for Faster Language Model Pre-Training
Jinbo Wang, Mingze Wang, Jiaqi Zhang +5
We propose GradPower, a lightweight gradient-transformation technique for accelerating language model pre-training. Given a gradient vector , GradPower first applies the…
Ditch the Denoiser: Emergence of Noise Robustness in Self-Supervised Learning from Data Curriculum
Wenquan Lu, Jiaqi Zhang, Hugues Van Assel +1
Self-Supervised Learning (SSL) has become a powerful solution to extract rich representations from unlabeled data. Yet, SSL research is mostly focused on clean, curated and high-qu…
Can Diffusion Models Disentangle? A Theoretical Perspective
Liming Wang, Muhammad Jehanzeb Mirza, Yishu Gong +6
This paper presents a novel theoretical framework for understanding how diffusion models can learn disentangled representations. Within this framework, we establish identifiability…
Condense, Don't Just Prune: Enhancing Efficiency and Performance in MoE Layer Pruning
Mingyu Cao, Gen Li, Jie Ji +6
Mixture-of-Experts (MoE) has garnered significant attention for its ability to scale up neural networks while utilizing the same or even fewer active parameters. However, MoE does…