Guangxin He

Pengbo Li, Feiyuan Zhang, Guangming Sheng et al.

GRPO- and PPO-style LLM post-training commonly sample multiple trajectories from the same prompt and then train on the resulting group. In long-context RL workloads, this shared prompt-side prefix can contain retrieved passages, visual tokens, tool schemas, system instructions, or task context, while the full rollout group is still too large to pack into one training microbatch. Standard dense trainers therefore recompute the same prefix forward and backward for every trajectory. We present a schedule-level reuse mechanism that decouples prefix and suffix computation. The schedule runs prefix forward once, executes suffixes as ordinary microbatches while reading prefix K/V and accumulating prefix-side gK/gV , and then runs prefix backward once on the accumulated gradient cache. This reordered schedule is equivalent to baseline training over real arithmetic and aligns numerically within finite-precision tolerance. Because only K/V and gK/gV are hot during suffix computation, the approach offloads dormant prefix activations, integrates with TP/EP/CP/PP and DP-style placement at the execution level, and preserves aux-loss-based MoE router semantics through logical prefix-token accounting. On dense Llama3-8B, Qwen3-8B, and MoE Qwen3-MoE-30B-A3B configurations, the schedule matches optimizer updates across TP/CP/PP/EP combinations, aligns on a 100-step real RL trace replay, reaches up to 4.395x speedup (2.930x under a conservative compile-on comparison) as prefix ratio and rollout group size grow, and reduces Phase-B peak HBM by up to 59.1%, extending the Llama3-8B capacity frontier from 17,920 to 29,696 total tokens.

Zipeng Qiu, You Peng, Guangxin He et al.

The advent of large language models (LLMs) has unlocked great opportunities in complex data management tasks, particularly in question answering (QA) over complicated multi-table relational data. Despite significant progress, systematically evaluating LLMs on multi-table QA remains a critical challenge due to the inherent complexity of analyzing heterogeneous table structures and potential large scale of serialized relational data. Existing benchmarks primarily focus on single-table QA, failing to capture the intricacies of reasoning across multiple relational tables, as required in real-world domains such as finance, healthcare, and e-commerce. To address this gap, we present TQA-Bench, a new multi-table QA benchmark designed to evaluate the capabilities of LLMs in tackling complex QA tasks over relational data. Our benchmark incorporates diverse relational database instances sourced from real-world public datasets and introduces a flexible sampling mechanism to create tasks with varying multi-table context lengths, ranging from 8K to 64K tokens. To ensure robustness and reliability, we integrate symbolic extensions into the evaluation framework, enabling the assessment of LLM reasoning capabilities beyond simple data retrieval or probabilistic pattern matching. We systematically evaluate a range of LLMs, both open-source and closed-source, spanning model scales from 7 billion to 70 billion parameters. Our extensive experiments reveal critical insights into the performance of LLMs in multi-table QA, highlighting both challenges and opportunities for advancing their application in complex, data-driven environments. Our benchmark implementation and results are available at https://github.com/Relaxed-System-Lab/TQA-Bench.

Haofei Cui, Guangxin He, Juanzhen Sun et al.

Short-range prediction of convective precipitation from weather radar observations is essential for severe weather warnings. However, deep learning models trained with pixel-wise error metrics tend to produce overly smooth forecasts that suppress intense echoes critical for hazard detection. This issue is exacerbated by insufficient multi-scale feature interaction and suboptimal fusion of heterogeneous geophysical inputs. We propose IMPA-Net (Integrated Multi-scale Predictive Attention Network), a deterministic 0-2 hour nowcasting framework that addresses these limitations through meteorologically-informed designs at the input, architecture, and loss function levels. A parameter-free Spatial Mixer reorganizes heterogeneous input channels at the mesoscale-$γ$ neighborhood (~2 km) via deterministic channel permutation, providing a structured cross-field prior. An integrated multi-scale predictive attention module serves as the spatiotemporal translator, capturing dynamics from mesoscale-$β$ to mesoscale-$γ$ scales. A Meteorologically-Aware Dynamic Loss employs three-level asymmetric weighting -- adapting across training epochs, storm intensity, and forecast lead time -- to counteract regression-to-the-mean. Evaluated against seven baselines on a multi-source radar dataset over eastern China, IMPA-Net raises the Heidke Skill Score at $\geq$45 dBZ from 0.049 (SimVP baseline) to 0.143 under matched settings. Relative to pySTEPS, it provides a better trade-off between severe-event detection and false-alarm control. Spectral analysis confirms preserved energy across mesoscale bands where competing methods show progressive smoothing. These improvements are shown within a single domain and convective regime; generalizability to other orographic and climatic regions remains to be tested.

Tianyi Bai, Zengjie Hu, Fupeng Sun et al.

Multi-modal large language models (MLLMs) have achieved remarkable capabilities by integrating visual perception with language understanding, enabling applications such as image-grounded dialogue, visual question answering, and scientific analysis. However, most MLLMs adopt a static inference paradigm, encoding the entire image into fixed visual tokens upfront, which limits their ability to iteratively refine understanding or adapt to context during inference. This contrasts sharply with human perception, which is dynamic, selective, and feedback-driven. In this work, we introduce a novel framework for inference-time visual token scaling that enables MLLMs to perform iterative, verifier-guided reasoning over visual content. We formulate the problem as a Markov Decision Process, involving a reasoner that proposes visual actions and a verifier, which is trained via multi-step Direct Preference Optimization (DPO), that evaluates these actions and determines when reasoning should terminate. To support this, we present a new dataset, VTS, comprising supervised reasoning trajectories (VTS-SFT) and preference-labeled reasoning comparisons (VTS-DPO). Our method significantly outperforms existing approaches across diverse visual reasoning benchmarks, offering not only improved accuracy but also more interpretable and grounded reasoning processes. These results demonstrate the promise of dynamic inference mechanisms for enabling fine-grained, context-aware visual reasoning in next-generation MLLMs.

Guangxin He, Shen Nie, Fengqi Zhu et al.

Diffusion LLMs have attracted growing interest, with plenty of recent work emphasizing their great potential in various downstream tasks; yet the long-context behavior of diffusion LLMs remains largely uncharted. We present a case study of post-training techniques for extending the context window of diffusion LLMs (i.e., LLaDA) without retraining from scratch. We show that a simple modification to the standard Rotary Positional Embeddings (RoPE) extension effectively accommodates the probabilistic modeling inherent in the diffusion process, enabling stable scaling to longer context ranges. We further compare masking strategies used during post-training and analyze their impact on optimization stability and long-range recall. Instantiating these insights, we introduce UltraLLaDA, a diffusion LLM with a 128K-token context window that, in our empirical evaluation on long-context tasks, significantly outperforms training-free baselines. Our experimental results highlight the special positional extension as a key lever for scaling diffusion LLMs to extended contexts and offer practical guidance for practitioners seeking 128K-scale context via efficient post-training.

Guangxin He, Yuan Cao, Yutong He et al.

Decentralized training of large language models offers the opportunity to pool computational resources across geographically distributed participants but faces significant network communication bottlenecks, particularly in pipeline-parallel settings. While pipeline parallelism partitions model layers across devices to handle large-scale models, it necessitates frequent communication of intermediate activations, creating challenges when network bandwidth is limited. Existing activation compression methods, such as AQ-SGD, mitigate quantization-induced errors through error compensation but impose prohibitive memory overhead by requiring storage of previous activations. To address these issues, we introduce TAH-Quant (Tile-wise Adaptive Hadamard Quantization), a novel activation quantization framework designed specifically for pipeline parallelism. Our approach integrates fine-grained tile-wise quantization for precise control, entropy-guided token-level adaptive bit allocation for optimal bit usage, and a Hadamard-based transform with pivot element swapping to effectively suppress quantization outliers. We further provide a theoretical analysis, proving that pipeline parallel training equipped with TAH-Quant maintains a convergence rate of $\mathcal{O}(1/\sqrt{T})$, matching that of vanilla stochastic gradient descent. Extensive experiments on diverse LLM tasks demonstrate that TAH-Quant achieves aggressive activation quantization (3-4 bits) ratio, which provides up to 4.3$\times$ end-to-end speedup without compromising training convergence, matches state-of-the-art methods, incurs no extra memory overhead, and generalizes well across different training scenarios.