arXiv subjects

Guoliang Zhao

Publications and source records attributed to Guoliang Zhao.

6 recordsLinked to original sources

arXiv2026

AngelSpec: Towards Real-World High Performance Inference with Speculative Decoding

Speculative decoding accelerates large language model inference without changing the target distribution, but no single drafting structure performs best across real-world workloads. Autoregressive multi-token prediction (MTP) is a lightweight, stable proposal mechanism, whereas block-parallel diffusion amortizes drafting latency over much longer candidate sequences; the better choice depends strongly on the output distribution. We present AngelSpec, a unified training framework for MTP and block-parallel speculative decoding that addresses this heterogeneity at three levels. At the training level, rather than fitting one universal drafter to a uniform data mixture, we co-specialize structure and data: the MTP drafter is trained on diverse conversational data for high-entropy open-ended chat, and the block-diffusion drafter on code and mathematics data for longer predictable continuations. At the architecture level, we propose DFly, a block-diffusion framework combining a hybrid target-conditioning backbone with a predecessor-conditioned autoregressive head, improving target-feature utilization and intra-block dependency modeling while keeping generation parallel. At the inference level, both acceptance length and verification cost vary with domain, request, online load, and hardware, so DFly treats verification as a shared batch-level resource: it reallocates compute toward high-confidence prefixes across requests and combines expected utility with a profiled cost model to adapt verification depth online. Across the Hy3 series, DFly raises the average accepted length on Hy3-A21B by roughly 30% and attains the highest average throughput at every tested concurrency from 4 to 64, a 1.98-2.40x speedup over autoregressive decoding and 10.5-11.8% higher throughput than DFlash. We release AngelSpec to support training and extending these methods.

cs.CL↗

arXiv2026

A Decomposition Perspective to Long-context Reasoning for LLMs

Long-context reasoning is essential for complex real-world applications, yet remains a significant challenge for Large Language Models (LLMs). Despite the rapid evolution in long-context reasoning, current research often overlooks the internal complexity of the long-context reasoning task itself. In this paper, we move beyond this holistic view and decompose long-context reasoning into a set of fundamental atomic skills, and we then automatically synthesize a suite of pseudo datasets, each explicitly targeting a specific atomic skill. Our empirical analysis confirms that proficiency in these atomic skills is strongly correlated with general long-text reasoning performance. Building on this insight, we employ reinforcement learning on these pseudo datasets to sharpen the model's atomic skills, in the hope of boosting its general long-context reasoning ability. Extensive experiments across multiple benchmarks demonstrate the effectiveness of our approach: it outperforms a strong baseline by an average margin of 7.7\% (improving from 46.3\% to 54.0\%) across Loogle, Loong, LongBench-v2, BrowscompLong, Ruler-qa2, and MRCR.

cs.CL↗

arXiv2026

Self-Distillation for Multi-Token Prediction

As Large Language Models (LLMs) scale up, inference efficiency becomes a critical bottleneck. Multi-Token Prediction (MTP) could accelerate LLM inference by predicting multiple future tokens in parallel. However, existing MTP approaches still face two challenges: limited acceptance rates of MTP heads, and difficulties in jointly training multiple MTP heads. Therefore, we propose MTP-D, a simple yet effective self-distillation method with minimal additional training cost, which boosts MTP head acceptance rates (+7.5\%) while maximumly preserving main-head performance. We also introduce a looped extension strategy for MTP-D, enabling effective and economical MTP head extension and further significant inference speedup to 1-head MTP (+220.4\%). Moreover, we systematically explore and validate key insights on the distillation strategies and the potential scalability of MTP through extensive experiments on seven benchmarks. These results demonstrate that our MTP-D and looped extension strategy effectively enhance MTP-head performance and inference efficiency, facilitating the practical usage of MTP in LLMs.

cs.CL↗

arXiv2026

Probing How Scalable Table Data Enhances General Long-Context Reasoning

As real-world tasks grow increasingly complex, long-context reasoning has become a core capability for Large Language Models (LLMs). However, few studies explore which data types are effective for long-context reasoning and why. We find that structured table data with periodic structures shows strong potential for long-context reasoning. Motivated by this observation, we mathematically analyze tabular dependency structures using mutual information, revealing periodic non-vanishing dependencies in table data. Furthermore, we systematically analyze the capabilities of structured table data, conduct relevant scaling experiments, and validate its underlying mechanisms for enhancing long-context reasoning, yielding several meaningful insights. Leveraging these insights, we propose a simple yet scalable pipeline(TableLong) for synthesizing high-quality, diverse, and verifiable structured table data to boost long-context reasoning via RL. Extensive experimental results demonstrate that table data significantly enhances the long-context reasoning capability of LLMs across multiple long-context benchmarks (+8.24\% on average), and even improves performance on out-of-domain benchmarks (+8.06\% on average). We hope that our insights provide practical guidance for effective post-training data to enhance long-context reasoning in LLMs.

cs.CL↗

arXiv2025

Towards a Comprehensive Scaling Law of Mixture-of-Experts

Mixture-of-Experts (MoE) models have become the consensus approach for enabling parameter-efficient scaling and cost-effective deployment in large language models. However, existing scaling laws for dense models are inapplicable to MoE models, which stems from three critical challenges: the multiplicity of influencing factors, their intricate coupling relationships and the non-monotonic nature of their performance impacts. They collectively necessitate a fine-grained investigation into MoE-specific scaling laws. In this work, we perform a systematic decomposition of MoE settings, identifying five key factors that influence model performance from both size and structural perspectives (data size ($D$), total model size ($N$), activated model size ($N_a$), number of active experts ($G$) and the ratio of shared experts ($S$)). Specifically, we design $446$ controlled experiments to characterize their marginal effects, ultimately constructing a comprehensive and precise joint MoE scaling law that considers all essential factors. Furthermore, we derive the theoretically optimal and practically efficiency-aware optimal configurations for $G$, $S$ and $N_a/N$ with detailed analyses. Our results demonstrate that the optimal settings for $G$ and $S$ are independent of both the model architecture and data size. With the scaling of $N$, the optimal activation parameter ratio of $N_a/N$ becomes sparser. Our proposed MoE scaling law could function as an accurate and insightful guidance to facilitate future MoE model design and training.

cs.LG↗

arXiv2024

Cross-Domain Transfer Learning Method for Thermal Adaptive Behavior Recognition with WiFi

A reliable comfort model is essential to improve occupant satisfaction and reduce building energy consumption. As two types of the most common and intuitive thermal adaptive behaviors, precise recognition of dressing and undressing can effectively support thermal comfort prediction. However, traditional activity recognition suffers from shortcomings in privacy, cost, and performance. To address the above issues, this study proposes a cross-domain transfer learning method for human dressing and undressing adaptive behavior recognition with WiFi. First, we determine the activity interval by calculating the sliding variance for denoised WiFi signals. Subsequently, short-time Fourier transform and discrete wavelet transform are performed to extract action information on the basis of time-frequency analysis. Ultimately, an efficient 1D CNN pre-trained model is integrated with the SVM algorithm as a hybrid model to enhance the identification robustness in new scenarios. Experiment results show that the hybrid model based on transfer learning provides a more accurate prediction for the adaptative behavior of target subjects, achieving 96.9% and 94.9% accuracy in two cases, respectively.

cs.HC↗