arXiv subjects

Xing Li

Publications and source records attributed to Xing Li.

At least 55 records · Page 3Linked to original sources

arXiv2025

Scaling Up, Speeding Up: A Benchmark of Speculative Decoding for Efficient LLM Test-Time Scaling

Test-time scaling has emerged as a powerful paradigm for enhancing the reasoning capabilities of large language models (LLMs) by allocating additional computational resources during inference. However, this paradigm is inherently inefficient due to the generation of redundant and repetitive reasoning traces, leading to significant computational overhead. Speculative decoding offers a promising avenue for mitigating this inefficiency, yet its efficacy in the structured, repetition-rich context of test-time scaling remains largely unexplored. To bridge this gap, we introduce the first comprehensive benchmark designed to evaluate speculative decoding methods for accelerating LLM test-time scaling. Our benchmark provides consistent experimental protocols across representative test-time scaling paradigms (e.g., Best-of-N sampling and multi-round thinking), enabling a fair comparison of three major categories of speculative decoding: model-based, training-based, and n-gram-based methods. Extensive experiments reveal that simple n-gram-based methods effectively capture repetitive patterns, demonstrating unique potential in accelerating test-time scaling. This phenomenon demonstrates the value of integrating n-gram-based methods with model-based or training-based approaches to balance acceleration for both repetitive and diverse reasoning in test-time scaling. We hope this benchmark spurs further research on speculative decoding for test-time scaling, enabling faster and more practical reasoning in LLMs through better handling of repetitive and diverse reasoning paths.

cs.CL↗

arXiv2025

ELF: Efficient Logic Synthesis by Pruning Redundancy in Refactoring

In electronic design automation, logic optimization operators play a crucial role in minimizing the gate count of logic circuits. However, their computation demands are high. Operators such as refactor conventionally form iterative cuts for each node, striving for a more compact representation - a task which often fails 98% on average. Prior research has sought to mitigate computational cost through parallelization. In contrast, our approach leverages a classifier to prune unsuccessful cuts preemptively, thus eliminating unnecessary resynthesis operations. Experiments on the refactor operator using the EPFL benchmark suite and 10 large industrial designs demonstrate that this technique can speedup logic optimization by 3.9x on average compared with the state-of-the-art ABC implementation.

cs.LG↗

arXiv2025

20-GHz bandwidth optical activation function based on a semiconductor laser

Optical neural networks usually execute the linear multiply-accumulate operation in the optical domain, whereas the nonlinear activation function is mostly implemented in the digital or electrical domain. Here we demonstrate a broadband activation function operated in the optical domain. The optical activation function (OAF) is achieved by a semiconductor laser with optical injection, which is operated in the unstable-locked regime above the Hopf bifurcation. The operation bandwidth of the OAF reaches as high as 20 GHz, which is limited by the relaxation oscillation resonance frequency of the semiconductor laser. The OAF mimics a ReLU function for modulation frequencies below 5 GHz, while resembling an ELU function for frequencies above 5 GHz.

physics.optics↗

arXiv2025

Hybrid Near-Far Field 6D Movable Antenna Design Exploiting Directional Sparsity and Deep Learning

Six-dimensional movable antenna (6DMA) has been identified as a new disruptive technology for future wireless systems to support a large number of users with only a few antennas. However, the intricate relationships between the signal carrier wavelength and the transceiver region size lead to inaccuracies in traditional far-field 6DMA channel model, causing discrepancies between the model predictions and the hybrid-field channel characteristics in practical 6DMA systems, where users might be in the far-field region relative to the antennas on the same 6DMA surface, while simultaneously being in the near-field region relative to different 6DMA surfaces. Moreover, due to the high-dimensional channel and the coupled position and rotation constraints, the estimation of the 6DMA channel and the joint design of the 6DMA positions and rotations and the transmit beamforming at the base station (BS) incur extremely high computational complexity. To address these issues, we propose an efficient hybrid-field generalized 6DMA channel model, which accounts for planar-wave propagation within individual 6DMA surfaces and spherical-wave propagation among different 6DMA surfaces. Furthermore, by leveraging directional sparsity, we propose a low-overhead channel estimation algorithm that efficiently constructs a complete channel map for all potential antenna position-rotation pairs while limiting the training overhead incurred by antenna movement. In addition, we propose a low-complexity design leveraging deep reinforcement learning (DRL), which facilitates the joint design of the 6DMA positions, rotations, and beamforming in a unified manner. Numerical results demonstrate that the proposed hybrid-field channel model and channel estimation algorithm outperform existing approaches and that the DRL-enhanced 6DMA system significantly surpasses flexible antenna systems.

cs.IT↗

arXiv2025

KAN-HyperpointNet for Point Cloud Sequence-Based 3D Human Action Recognition

Point cloud sequence-based 3D action recognition has achieved impressive performance and efficiency. However, existing point cloud sequence modeling methods cannot adequately balance the precision of limb micro-movements with the integrity of posture macro-structure, leading to the loss of crucial information cues in action inference. To overcome this limitation, we introduce D-Hyperpoint, a novel data type generated through a D-Hyperpoint Embedding module. D-Hyperpoint encapsulates both regional-momentary motion and global-static posture, effectively summarizing the unit human action at each moment. In addition, we present a D-Hyperpoint KANsMixer module, which is recursively applied to nested groupings of D-Hyperpoints to learn the action discrimination information and creatively integrates Kolmogorov-Arnold Networks (KAN) to enhance spatio-temporal interaction within D-Hyperpoints. Finally, we propose KAN-HyperpointNet, a spatio-temporal decoupled network architecture for 3D action recognition. Extensive experiments on two public datasets: MSR Action3D and NTU-RGB+D 60, demonstrate the state-of-the-art performance of our method.

cs.CV↗

arXiv2025

TrimR: Verifier-based Training-Free Thinking Compression for Efficient Test-Time Scaling

Large Reasoning Models (LRMs) demonstrate exceptional capability in tackling complex mathematical, logical, and coding tasks by leveraging extended Chain-of-Thought (CoT) reasoning. Test-time scaling methods, such as prolonging CoT with explicit token-level exploration, can push LRMs' accuracy boundaries, but they incur significant decoding overhead. A key inefficiency source is LRMs often generate redundant thinking CoTs, which demonstrate clear structured overthinking and underthinking patterns. Inspired by human cognitive reasoning processes and numerical optimization theories, we propose TrimR, a verifier-based, training-free, efficient framework for dynamic CoT compression to trim reasoning and enhance test-time scaling, explicitly tailored for production-level deployment. Our method employs a lightweight, pretrained, instruction-tuned verifier to detect and truncate redundant intermediate thoughts of LRMs without any LRM or verifier fine-tuning. We present both the core algorithm and asynchronous online system engineered for high-throughput industrial applications. Empirical evaluations on Ascend NPUs and vLLM show that our framework delivers substantial gains in inference efficiency under large-batch workloads. In particular, on the four MATH500, AIME24, AIME25, and GPQA benchmarks, the reasoning runtime of Pangu Pro MoE, Pangu-R-38B, QwQ-32B, and DeepSeek-R1-Distill-Qwen-32B is improved by up to 70% with negligible impact on accuracy.

cs.LG↗

arXiv2025

Unlocking Efficient Long-to-Short LLM Reasoning with Model Merging

The transition from System 1 to System 2 reasoning in large language models (LLMs) has marked significant advancements in handling complex tasks through deliberate, iterative thinking. However, this progress often comes at the cost of efficiency, as models tend to overthink, generating redundant reasoning steps without proportional improvements in output quality. Long-to-Short (L2S) reasoning has emerged as a promising solution to this challenge, aiming to balance reasoning depth with practical efficiency. While existing approaches, such as supervised fine-tuning (SFT), reinforcement learning (RL), and prompt engineering, have shown potential, they are either computationally expensive or unstable. Model merging, on the other hand, offers a cost-effective and robust alternative by integrating the quick-thinking capabilities of System 1 models with the methodical reasoning of System 2 models. In this work, we present a comprehensive empirical study on model merging for L2S reasoning, exploring diverse methodologies, including task-vector-based, SVD-based, and activation-informed merging. Our experiments reveal that model merging can reduce average response length by up to 55% while preserving or even improving baseline performance. We also identify a strong correlation between model scale and merging efficacy with extensive evaluations on 1.5B/7B/14B/32B models. Furthermore, we investigate the merged model's ability to self-critique and self-correct, as well as its adaptive response length based on task complexity. Our findings highlight model merging as a highly efficient and effective paradigm for L2S reasoning, offering a practical solution to the overthinking problem while maintaining the robustness of System 2 reasoning. This work can be found on Github https://github.com/hahahawu/Long-to-Short-via-Model-Merging.

cs.CL↗

arXiv2025

Recurrence relations for the generalized Laguerre and Charlier orthogonal polynomials and discrete Painlevé equations on the $D_{6}^{(1)}$ Sakai surface

This paper concerns the discrete version of the Painlevé identification problem, i.e., how to recognize a certain recurrence relation as a discrete Painlevé equation. Often some clues can be seen from the setting of the problem, e.g., when the recurrence is connected with some differential Painlevé equation, or from the geometry of the configuration of indeterminate points of the equation. The main message of our paper is that, in fact, this only allows us to identify the configuration space of the dynamic system, but not the dynamics themselves. The refined version of the identification problem lies in determining, up to the conjugation, the translation direction of the dynamics, which in turn requires the full power of the geometric theory of Painlevé equations. To illustrate this point, in this paper we consider two examples of such recurrences that appear in the theory of orthogonal polynomials. We choose these examples because they get regularized on the same family of Sakai surfaces, but at the same time are not equivalent, since they result in non-equivalent translation directions. In addition, we show the effectiveness of a recently proposed identification procedure for discrete Painlevé equations using Sakai's geometric approach for answering such questions. In particular, this approach requires no a priori knowledge of a possible type of the equation.

nlin.SI↗

arXiv2025

Pig behavior dataset and Spatial-temporal perception and enhancement networks based on the attention mechanism for pig behavior recognition

The recognition of pig behavior plays a crucial role in smart farming and welfare assurance for pigs. Currently, in the field of pig behavior recognition, the lack of publicly available behavioral datasets not only limits the development of innovative algorithms but also hampers model robustness and algorithm optimization.This paper proposes a dataset containing 13 pig behaviors that significantly impact welfare.Based on this dataset, this paper proposes a spatial-temporal perception and enhancement networks based on the attention mechanism to model the spatiotemporal features of pig behaviors and their associated interaction areas in video data. The network is composed of a spatiotemporal perception network and a spatiotemporal feature enhancement network. The spatiotemporal perception network is responsible for establishing connections between the pigs and the key regions of their behaviors in the video data. The spatiotemporal feature enhancement network further strengthens the important spatial features of individual pigs and captures the long-term dependencies of the spatiotemporal features of individual behaviors by remodeling these connections, thereby enhancing the model's perception of spatiotemporal changes in pig behaviors. Experimental results demonstrate that on the dataset established in this paper, our proposed model achieves a MAP score of 75.92%, which is an 8.17% improvement over the best-performing traditional model. This study not only improces the accuracy and generalizability of individual pig behavior recognition but also provides new technological tools for modern smart farming. The dataset and related code will be made publicly available alongside this paper.

cs.CV↗

arXiv2025

Ionization phase-shifts between two laser-dressed states

Resonance-enhanced multiphoton ionization (REMPI) in potassium atom is investigated within the strong coupling regime using photoelectron momentum imaging techniques. The kinetic energy distribution of the ionized electrons reveals the eigenenergies of the dressed states, which exhibit Autler-Townes (AT) splitting. This splitting is proportional to the laser field strength. Partial wave analysis of the photoelectrons angular distributions (PAD) uncovers relative phase shifts between the dressed states, shedding light on attosecond ionization time delays. The observed phase shift between the two electron wave packets generated by the AT splitting arises from the combined effects of the Coulomb phase and the quantum defect phase.

physics.atom-ph↗

arXiv2025

Algebro-geometric integration to the discrete Chen-Lee-Liu system

Algebro-geometric solutions for the discrete Chen-Lee-Liu (CLL) system are derived in this paper. We construct a nonlinear integrable symplectic map which is used to define discrete phase flows. Compatibility of the maps with different parameters gives rise to the discrete CLL system whose solutions (discrete potentials) can be formulated through the discrete phase flows. Baker-Akhiezer functions are introduced and their asymptotic behaviors are analyzed. Consequently, we are able to reconstruct the discrete potentials in terms of the Riemann theta functions. These results can be extended to 3-dimensional case and algebro-geometric solutions of the discrete modified Kadomtsev-Petviashvili equation are obtained. Some solutions of genus one case are illustrated.

nlin.SI↗

arXiv2025

Logic Synthesis Optimization with Predictive Self-Supervision via Causal Transformers

Contemporary hardware design benefits from the abstraction provided by high-level logic gates, streamlining the implementation of logic circuits. Logic Synthesis Optimization (LSO) operates at one level of abstraction within the Electronic Design Automation (EDA) workflow, targeting improvements in logic circuits with respect to performance metrics such as size and speed in the final layout. Recent trends in the field show a growing interest in leveraging Machine Learning (ML) for EDA, notably through ML-guided logic synthesis utilizing policy-based Reinforcement Learning (RL) methods.Despite these advancements, existing models face challenges such as overfitting and limited generalization, attributed to constrained public circuits and the expressiveness limitations of graph encoders. To address these hurdles, and tackle data scarcity issues, we introduce LSOformer, a novel approach harnessing Autoregressive transformer models and predictive SSL to predict the trajectory of Quality of Results (QoR). LSOformer integrates cross-attention modules to merge insights from circuit graphs and optimization sequences, thereby enhancing prediction accuracy for QoR metrics. Experimental studies validate the effectiveness of LSOformer, showcasing its superior performance over baseline architectures in QoR prediction tasks, where it achieves improvements of 5.74%, 4.35%, and 17.06% on the EPFL, OABCD, and proprietary circuits datasets, respectively, in inductive setup.

cs.AI↗

arXiv2025

Circuit Transformer: A Transformer That Preserves Logical Equivalence

Implementing Boolean functions with circuits consisting of logic gates is fundamental in digital computer design. However, the implemented circuit must be exactly equivalent, which hinders generative neural approaches on this task due to their occasionally wrong predictions. In this study, we introduce a generative neural model, the "Circuit Transformer", which eliminates such wrong predictions and produces logic circuits strictly equivalent to given Boolean functions. The main idea is a carefully designed decoding mechanism that builds a circuit step-by-step by generating tokens, which has beneficial "cutoff properties" that block a candidate token once it invalidate equivalence. In such a way, the proposed model works similar to typical LLMs while logical equivalence is strictly preserved. A Markov decision process formulation is also proposed for optimizing certain objectives of circuits. Experimentally, we trained an 88-million-parameter Circuit Transformer to generate equivalent yet more compact forms of input circuits, outperforming existing neural approaches on both synthetic and real world benchmarks, without any violation of equivalence constraints.

cs.LG↗

arXiv2024

HiCat: A Semi-Supervised Approach for Cell Type Annotation

We introduce HiCat (Hybrid Cell Annotation using Transformative embeddings), a novel semi-supervised pipeline for annotating cell types from single-cell RNA sequencing data. HiCat fuses the strengths of supervised learning for known cell types with unsupervised learning to identify novel types. This hybrid approach incorporates both reference and query genomic data for feature engineering, enhancing the embedding learning process, increasing the effective sample size for unsupervised techniques, and improving the transferability of the supervised model trained on reference data when applied to query datasets. The pipeline follows six key steps: (1) removing batch effects using Harmony to generate a 50-dimensional principal component embedding; (2) applying UMAP for dimensionality reduction to two dimensions to capture crucial data patterns; (3) conducting unsupervised clustering of cells with DBSCAN, yielding a one-dimensional cluster membership vector; (4) merging the multi-resolution results of the previous steps into a 53-dimensional feature space that encompasses both reference and query data; (5) training a CatBoost model on the reference dataset to predict cell types in the query dataset; and (6) resolving inconsistencies between the supervised predictions and unsupervised cluster labels. When benchmarked on 10 publicly available genomic datasets, HiCat surpasses other methods, particularly in differentiating and identifying multiple new cell types. Its capacity to accurately classify novel cell types showcases its robustness and adaptability within intricate biological datasets.

q-bio.BM↗

arXiv2024

Constructive RNNs: An Error-Recurrence Perspective on Time-Variant Zero Finding Problem Solving Under Uncertainty

When facing time-variant problems in analog computing, the desirable RNN design requires finite-time convergence and robustness with respect to various types of uncertainties, due to the time-variant nature and difficulties in implementation. It is very worthwhile to explore terminal zeroing neural networks, through examining and applying available attracting laws. In this paper, from a control-theoretic point of view, an error recurrence system approach is presented by equipping with uncertainty compensation in the pre-specified error dynamics, capable of enhancing robustness properly. Novel rectifying actions are designed to make finite-time settling so that the convergence speed and the computing accuracy of time-variant computing can be improved. Double-power and power-exponential rectifying actions are respectively formed to construct specific models, while the particular expressions of settling time function for the former are presented, and for the latter the proximate settling-time estimations are given, with which the fixed-time convergence of the corresponding models is in turn established. Moreover, the uncertainty compensation by the signum/smoothing-signum techniques are adopted for finite-duration stabilization. Theoretical results are presented to demonstrate effectiveness (involving fixed-time convergence and robustness) of the proposed computing schemes for the time-variant QP problem solving.

eess.SY↗

arXiv2024

MTLSO: A Multi-Task Learning Approach for Logic Synthesis Optimization

Electronic Design Automation (EDA) is essential for IC design and has recently benefited from AI-based techniques to improve efficiency. Logic synthesis, a key EDA stage, transforms high-level hardware descriptions into optimized netlists. Recent research has employed machine learning to predict Quality of Results (QoR) for pairs of And-Inverter Graphs (AIGs) and synthesis recipes. However, the severe scarcity of data due to a very limited number of available AIGs results in overfitting, significantly hindering performance. Additionally, the complexity and large number of nodes in AIGs make plain GNNs less effective for learning expressive graph-level representations. To tackle these challenges, we propose MTLSO - a Multi-Task Learning approach for Logic Synthesis Optimization. On one hand, it maximizes the use of limited data by training the model across different tasks. This includes introducing an auxiliary task of binary multi-label graph classification alongside the primary regression task, allowing the model to benefit from diverse supervision sources. On the other hand, we employ a hierarchical graph representation learning strategy to improve the model's capacity for learning expressive graph-level representations of large AIGs, surpassing traditional plain GNNs. Extensive experiments across multiple datasets and against state-of-the-art baselines demonstrate the superiority of our method, achieving an average performance gain of 8.22\% for delay and 5.95\% for area.

cs.LG↗

arXiv2024

SeaDAG: Semi-autoregressive Diffusion for Conditional Directed Acyclic Graph Generation

We introduce SeaDAG, a semi-autoregressive diffusion model for conditional generation of Directed Acyclic Graphs (DAGs). Considering their inherent layer-wise structure, we simulate layer-wise autoregressive generation by designing different denoising speed for different layers. Unlike conventional autoregressive generation that lacks a global graph structure view, our method maintains a complete graph structure at each diffusion step, enabling operations such as property control that require the full graph structure. Leveraging this capability, we evaluate the DAG properties during training by employing a graph property decoder. We explicitly train the model to learn graph conditioning with a condition loss, which enhances the diffusion model's capacity to generate graphs that are both realistic and aligned with specified properties. We evaluate our method on two representative conditional DAG generation tasks: (1) circuit generation from truth tables, where precise DAG structures are crucial for realizing circuit functionality, and (2) molecule generation based on quantum properties. Our approach demonstrates promising results, generating high-quality and realistic DAGs that closely align with given conditions.

cs.LG↗

arXiv2024

RTLRewriter: Methodologies for Large Models aided RTL Code Optimization

Register Transfer Level (RTL) code optimization is crucial for enhancing the efficiency and performance of digital circuits during early synthesis stages. Currently, optimization relies heavily on manual efforts by skilled engineers, often requiring multiple iterations based on synthesis feedback. In contrast, existing compiler-based methods fall short in addressing complex designs. This paper introduces RTLRewriter, an innovative framework that leverages large models to optimize RTL code. A circuit partition pipeline is utilized for fast synthesis and efficient rewriting. A multi-modal program analysis is proposed to incorporate vital visual diagram information as optimization cues. A specialized search engine is designed to identify useful optimization guides, algorithms, and code snippets that enhance the model ability to generate optimized RTL. Additionally, we introduce a Cost-aware Monte Carlo Tree Search (C-MCTS) algorithm for efficient rewriting, managing diverse retrieved contents and steering the rewriting results. Furthermore, a fast verification pipeline is proposed to reduce verification cost. To cater to the needs of both industry and academia, we propose two benchmarking suites: the Large Rewriter Benchmark, targeting complex scenarios with extensive circuit partitioning, optimization trade-offs, and verification challenges, and the Small Rewriter Benchmark, designed for a wider range of scenarios and patterns. Our comparative analysis with established compilers such as Yosys and E-graph demonstrates significant improvements, highlighting the benefits of integrating large models into the early stages of circuit design. We provide our benchmarks at https://github.com/yaoxufeng/RTLRewriter-Bench.

cs.AR↗