arXiv subjects

Hongwei Liu

Publications and source records attributed to Hongwei Liu.

At least 55 records · Page 3Linked to original sources

arXiv2025

Large Processor Chip Model

Computer System Architecture serves as a crucial bridge between software applications and the underlying hardware, encompassing components like compilers, CPUs, coprocessors, and RTL designs. Its development, from early mainframes to modern domain-specific architectures, has been driven by rising computational demands and advancements in semiconductor technology. However, traditional paradigms in computer system architecture design are confronting significant challenges, including a reliance on manual expertise, fragmented optimization across software and hardware layers, and high costs associated with exploring expansive design spaces. While automated methods leveraging optimization algorithms and machine learning have improved efficiency, they remain constrained by a single-stage focus, limited data availability, and a lack of comprehensive human domain knowledge. The emergence of large language models offers transformative opportunities for the design of computer system architecture. By leveraging the capabilities of LLMs in areas such as code generation, data analysis, and performance modeling, the traditional manual design process can be transitioned to a machine-based automated design approach. To harness this potential, we present the Large Processor Chip Model (LPCM), an LLM-driven framework aimed at achieving end-to-end automated computer architecture design. The LPCM is structured into three levels: Human-Centric; Agent-Orchestrated; and Model-Governed. This paper utilizes 3D Gaussian Splatting as a representative workload and employs the concept of software-hardware collaborative design to examine the implementation of the LPCM at Level 1, demonstrating the effectiveness of the proposed approach. Furthermore, this paper provides an in-depth discussion on the pathway to implementing Level 2 and Level 3 of the LPCM, along with an analysis of the existing challenges.

cs.AR↗

arXiv2025

Deciphering Trajectory-Aided LLM Reasoning: An Optimization Perspective

We propose a novel framework for comprehending the reasoning capabilities of large language models (LLMs) through the perspective of meta-learning. By conceptualizing reasoning trajectories as pseudo-gradient descent updates to the LLM's parameters, we identify parallels between LLM reasoning and various meta-learning paradigms. We formalize the training process for reasoning tasks as a meta-learning setup, with each question treated as an individual task, and reasoning trajectories serving as the inner loop optimization for adapting model parameters. Once trained on a diverse set of questions, the LLM develops fundamental reasoning capabilities that can generalize to previously unseen questions. Extensive empirical evaluations substantiate the strong connection between LLM reasoning and meta-learning, exploring several issues of significant interest from a meta-learning standpoint. Our work not only enhances the understanding of LLM reasoning but also provides practical insights for improving these models through established meta-learning techniques.

cs.CL↗

arXiv2025

Beyond Matryoshka: Revisiting Sparse Coding for Adaptive Representation

Many large-scale systems rely on high-quality deep representations (embeddings) to facilitate tasks like retrieval, search, and generative modeling. Matryoshka Representation Learning (MRL) recently emerged as a solution for adaptive embedding lengths, but it requires full model retraining and suffers from noticeable performance degradations at short lengths. In this paper, we show that sparse coding offers a compelling alternative for achieving adaptive representation with minimal overhead and higher fidelity. We propose Contrastive Sparse Representation (CSR), a method that sparsifies pre-trained embeddings into a high-dimensional but selectively activated feature space. By leveraging lightweight autoencoding and task-aware contrastive objectives, CSR preserves semantic quality while allowing flexible, cost-effective inference at different sparsity levels. Extensive experiments on image, text, and multimodal benchmarks demonstrate that CSR consistently outperforms MRL in terms of both accuracy and retrieval speed-often by large margins-while also cutting training time to a fraction of that required by MRL. Our results establish sparse coding as a powerful paradigm for adaptive representation learning in real-world applications where efficiency and fidelity are both paramount. Code is available at https://github.com/neilwen987/CSR_Adaptive_Rep

cs.LG↗

arXiv2025

Discovering Fine-Grained Visual-Concept Relations by Disentangled Optimal Transport Concept Bottleneck Models

Concept Bottleneck Models (CBMs) try to make the decision-making process transparent by exploring an intermediate concept space between the input image and the output prediction. Existing CBMs just learn coarse-grained relations between the whole image and the concepts, less considering local image information, leading to two main drawbacks: i) they often produce spurious visual-concept relations, hence decreasing model reliability; and ii) though CBMs could explain the importance of every concept to the final prediction, it is still challenging to tell which visual region produces the prediction. To solve these problems, this paper proposes a Disentangled Optimal Transport CBM (DOT-CBM) framework to explore fine-grained visual-concept relations between local image patches and concepts. Specifically, we model the concept prediction process as a transportation problem between the patches and concepts, thereby achieving explicit fine-grained feature alignment. We also incorporate orthogonal projection losses within the modality to enhance local feature disentanglement. To further address the shortcut issues caused by statistical biases in the data, we utilize the visual saliency map and concept label statistics as transportation priors. Thus, DOT-CBM can visualize inversion heatmaps, provide more reliable concept predictions, and produce more accurate class predictions. Comprehensive experiments demonstrate that our proposed DOT-CBM achieves SOTA performance on several tasks, including image classification, local part detection and out-of-distribution generalization.

cs.CV↗

arXiv2025

Trustworthy Enhanced Multi-view Multi-modal Alzheimer's Disease Prediction with Brain-wide Imaging Transcriptomics Data

Brain transcriptomics provides insights into the molecular mechanisms by which the brain coordinates its functions and processes. However, existing multimodal methods for predicting Alzheimer's disease (AD) primarily rely on imaging and sometimes genetic data, often neglecting the transcriptomic basis of brain. Furthermore, while striving to integrate complementary information between modalities, most studies overlook the informativeness disparities between modalities. Here, we propose TMM, a trusted multiview multimodal graph attention framework for AD diagnosis, using extensive brain-wide transcriptomics and imaging data. First, we construct view-specific brain regional co-function networks (RRIs) from transcriptomics and multimodal radiomics data to incorporate interaction information from both biomolecular and imaging perspectives. Next, we apply graph attention (GAT) processing to each RRI network to produce graph embeddings and employ cross-modal attention to fuse transcriptomics-derived embedding with each imagingderived embedding. Finally, a novel true-false-harmonized class probability (TFCP) strategy is designed to assess and adaptively adjust the prediction confidence of each modality for AD diagnosis. We evaluate TMM using the AHBA database with brain-wide transcriptomics data and the ADNI database with three imaging modalities (AV45-PET, FDG-PET, and VBM-MRI). The results demonstrate the superiority of our method in identifying AD, EMCI, and LMCI compared to state-of-the-arts. Code and data are available at https://github.com/Yaolab-fantastic/TMM.

cs.AI↗

arXiv2025

Synthesis-related nanoscale defects in Mo-based Janus monolayers revealed by cross-correlated AFM and TERS imaging

Two-dimensional (2D) Janus transition metal dichalcogenides (TMDs) are promising candidates for various applications in non-linear optics, energy harvesting, and catalysis. These materials are usually synthesized via chemical conversion of pristine TMDs. Nanometer-scale characterization of the obtained Janus materials' morphology and local composition is crucial for both the synthesis optimization and the future device applications. In this work, we present a cross-correlated atomic force microscopy (AFM) and tip-enhanced Raman spectroscopy (TERS) study of Janus $\mathrm{Mo}_{\mathrm{Se}}^{\mathrm{S}}$ and Janus $\mathrm{Mo}_{\mathrm{S}}^{\mathrm{Se}}$ monolayers synthesized by the hydrogen plasma-assisted chemical conversion of $\mathrm{MoSe}_2$ and $\mathrm{MoS}_2$, respectively. We demonstrate how the choice of the growth substrate and the starting TMD affects the morphology of the resulting Janus material. Furthermore, by employing TERS imaging, we demonstrate the presence of nanoscale islands (~20 nm across) of $\mathrm{MoSe}_2$-$\mathrm{Mo}_{\mathrm{Se}}^{\mathrm{S}}$ ($\mathrm{MoS}_2$-$\mathrm{Mo}_{\mathrm{S}}^{\mathrm{Se}}$) vertical heterostructures originating from the bilayer nanoislands in the precursor monolayer crystals. The understanding of the origins of nanoscale defects in Janus TMDs revealed in our study can help with further optimization of the Janus conversion process towards uniform and wrinkle-/crack-free Janus materials. Moreover, our work shows that cross-correlated AFM and TERS imaging is a powerful and accessible method for studying nanoscale composition and defects in Janus TMD monolayers.

cond-mat.mtrl-sci↗

arXiv2025

Low-Rank Adaptation of Pre-Trained Stable Diffusion for Rigid-Body Target ISAR Imaging

Traditional range-instantaneous Doppler (RID) methods for rigid-body target imaging often suffer from low resolution due to the limitations of time-frequency analysis (TFA). To address this challenge, our primary focus is on obtaining high resolution time-frequency representations (TFRs) from their low resolution counterparts. Recognizing that the curve features of TFRs are a specific type of texture feature, we argue that pre trained generative models such as Stable Diffusion (SD) are well suited for enhancing TFRs, thanks to their powerful capability in capturing texture representations. Building on this insight, we propose a novel inverse synthetic aperture radar (ISAR) imaging method for rigid-body targets, leveraging the low-rank adaptation (LoRA) of a pre-trained SD model. Our approach adopts the basic structure and pre-trained parameters of SD Turbo while incorporating additional linear operations for LoRA and adversarial training to achieve super-resolution and noise suppression. Then we integrate LoRA-SD into the RID-based ISAR imaging, enabling sharply focused and denoised imaging with super-resolution capabilities. We evaluate our method using both simulated and real radar data. The experimental results demonstrate the superiority of our approach in frequency es timation and ISAR imaging compared to traditional methods. Notably, the generalization capability is verified by training on simulated radar data and testing on measured radar data.

cs.CV↗

arXiv2025

Explaining Domain Shifts in Language: Concept erasing for Interpretable Image Classification

Concept-based models can map black-box representations to human-understandable concepts, which makes the decision-making process more transparent and then allows users to understand the reason behind predictions. However, domain-specific concepts often impact the final predictions, which subsequently undermine the model generalization capabilities, and prevent the model from being used in high-stake applications. In this paper, we propose a novel Language-guided Concept-Erasing (LanCE) framework. In particular, we empirically demonstrate that pre-trained vision-language models (VLMs) can approximate distinct visual domain shifts via domain descriptors while prompting large Language Models (LLMs) can easily simulate a wide range of descriptors of unseen visual domains. Then, we introduce a novel plug-in domain descriptor orthogonality (DDO) regularizer to mitigate the impact of these domain-specific concepts on the final predictions. Notably, the DDO regularizer is agnostic to the design of concept-based models and we integrate it into several prevailing models. Through evaluation of domain generalization on four standard benchmarks and three newly introduced benchmarks, we demonstrate that DDO can significantly improve the out-of-distribution (OOD) generalization over the previous state-of-the-art concept-based models.Our code is available at https://github.com/joeyz0z/LanCE.

cs.CV↗

arXiv2025

Multi-constraint Graph Partitioning Problems Via Recursive Bipartition Algorithm Based on Subspace Minimization Conjugate Gradient Method

The graph partitioning problem is a well-known NP-hard problem. In this paper, we formulate a 0-1 quadratic integer programming model for the graph partitioning problem with vertex weight constraints and fixed vertex constraints, and propose a recursive bipartition algorithm based on the subspace minimization conjugate gradient method. To alleviate the difficulty of solving the model, the constrained problem is transformed into an unconstrained optimization problem using equilibrium terms, elimination methods, and trigonometric properties, and solved via an accelerated subspace minimization conjugate gradient algorithm. Initial feasible partitions are generated using a hyperplane rounding algorithm, followed by heuristic refinement strategies, including one-neighborhood and two-interchange adjustments, to iteratively improve the results. Numerical experiments on knapsack-constrained graph partitioning and industrial examples demonstrate the effectiveness and feasibility of the proposed algorithm.

math.OC↗

arXiv2025

Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning

Reasoning abilities, especially those for solving complex math problems, are crucial components of general intelligence. Recent advances by proprietary companies, such as o-series models of OpenAI, have made remarkable progress on reasoning tasks. However, the complete technical details remain unrevealed, and the techniques that are believed certainly to be adopted are only reinforcement learning (RL) and the long chain of thoughts. This paper proposes a new RL framework, termed OREAL, to pursue the performance limit that can be achieved through \textbf{O}utcome \textbf{RE}w\textbf{A}rd-based reinforcement \textbf{L}earning for mathematical reasoning tasks, where only binary outcome rewards are easily accessible. We theoretically prove that behavior cloning on positive trajectories from best-of-N (BoN) sampling is sufficient to learn the KL-regularized optimal policy in binary feedback environments. This formulation further implies that the rewards of negative samples should be reshaped to ensure the gradient consistency between positive and negative samples. To alleviate the long-existing difficulties brought by sparse rewards in RL, which are even exacerbated by the partial correctness of the long chain of thought for reasoning tasks, we further apply a token-level reward model to sample important tokens in reasoning trajectories for learning. With OREAL, for the first time, a 7B model can obtain 94.0 pass@1 accuracy on MATH-500 through RL, being on par with 32B models. OREAL-32B also surpasses previous 32B models trained by distillation with 95.0 pass@1 accuracy on MATH-500. Our investigation also indicates the importance of initial policy models and training queries for RL. Code, models, and data will be released to benefit future research\footnote{https://github.com/InternLM/OREAL}.

cs.CL↗

arXiv2025

About the Rankin and Bergé-Martinet Constants from a Coding Theory View Point

The Rankin constant $γ_{n,l}$ measures the largest volume of the densest sublattice of rank $l$ of a lattice $Λ\in \RR^n$ over all such lattices of rank $n$. The Bergé-Martinet constant $γ'_{n,l}$ is a variation that takes into account the dual lattice. Exact values and bounds for both constants are mostly open in general. We consider the case of lattices built from linear codes, and look at bounds on $γ_{n,l}$ and $γ'_{n,l}$. In particular, we revisit known results for $n=3,4,5,8$ and give lower and upper bounds for the open cases $γ_{5,2},γ_{7,2}$ and $γ'_{5,2},γ'_{7,2}$.

cs.IT↗

arXiv2024

New Constructions of Optimal Binary LCD Codes

Linear complementary dual (LCD) codes can provide an optimum linear coding solution for the two-user binary adder channel. LCD codes also can be used to against side-channel attacks and fault non-invasive attacks. Let $d_{LCD}(n, k)$ denote the maximum value of $d$ for which a binary $[n,k, d]$ LCD code exists. In \cite{BS21}, Bouyuklieva conjectured that $d_{LCD}(n+1, k)=d_{LCD}(n, k)$ or $d_{LCD}(n, k) + 1$ for any lenth $n$ and dimension $k \ge 2$. In this paper, we first prove Bouyuklieva's conjecture \cite{BS21} by constructing a binary $[n,k,d-1]$ LCD codes from a binary $[n+1,k,d]$ $LCD_{o,e}$ code, when $d \ge 3$ and $k \ge 2$. Then we provide a distance lower bound for binary LCD codes by expanded codes, and use this bound and some methods such as puncturing, shortening, expanding and extension, we construct some new binary LCD codes. Finally, we improve some previously known values of $d_{LCD}(n, k)$ of lengths $38 \le n \le 40$ and dimensions $9 \le k \le 15$. We also obtain some values of $d_{LCD}(n, k)$ with $41 \le n \le 50$ and $6 \le k \le n-6$.

cs.IT↗

arXiv2024

Constructions of non-Generalized Reed-Solomon MDS codes

Generalized Reed-Solomon codes form the most prominent class of maximum distance separable (MDS) codes, codes that are optimal in the sense that their minimum distance cannot be improved for a given length and code size. The study of codes that are MDS yet not generalized Reed-Solomon codes, called non-generalized Reed-Solomon MDS codes, started with the work by Roth and Lemple (1989), where the first examples where exhibited. It then gained traction thanks to the work by Beelen (2017), who introduced twisted Reed-Solomon codes, and showed that families of such codes are non-generalized Reed-Solomon MDS codes. Finding non-generalized Reed-Solomon MDS codes is naturally motivated by the classification of MDS codes. In this paper, we provide a generic construction of MDS codes, yielding infinitely many examples. We then explicit families of non-generalized Reed-Solomon MDS codes. Finally we position some of the proposed codes with respect to generalized twisted Reed-Solomon codes, and provide new view points on this family of codes.

cs.IT↗

arXiv2024

MVKTrans: Multi-View Knowledge Transfer for Robust Multiomics Classification

The distinct characteristics of multiomics data, including complex interactions within and across biological layers and disease heterogeneity (e.g., heterogeneity in etiology and clinical symptoms), drive us to develop novel designs to address unique challenges in multiomics prediction. In this paper, we propose the multi-view knowledge transfer learning (MVKTrans) framework, which transfers intra- and inter-omics knowledge in an adaptive manner by reviewing data heterogeneity and suppressing bias transfer, thereby enhancing classification performance. Specifically, we design a graph contrastive module that is trained on unlabeled data to effectively learn and transfer the underlying intra-omics patterns to the supervised task. This unsupervised pretraining promotes learning general and unbiased representations for each modality, regardless of the downstream tasks. In light of the varying discriminative capacities of modalities across different diseases and/or samples, we introduce an adaptive and bi-directional cross-omics distillation module. This module automatically identifies richer modalities and facilitates dynamic knowledge transfer from more informative to less informative omics, thereby enabling a more robust and generalized integration. Extensive experiments on four real biomedical datasets demonstrate the superior performance and robustness of MVKTrans compared to the state-of-the-art. Code and data are available at https://github.com/Yaolab-fantastic/MVKTrans.

cs.LG↗

arXiv2024

CompassJudger-1: All-in-one Judge Model Helps Model Evaluation and Evolution

Efficient and accurate evaluation is crucial for the continuous improvement of large language models (LLMs). Among various assessment methods, subjective evaluation has garnered significant attention due to its superior alignment with real-world usage scenarios and human preferences. However, human-based evaluations are costly and lack reproducibility, making precise automated evaluators (judgers) vital in this process. In this report, we introduce \textbf{CompassJudger-1}, the first open-source \textbf{all-in-one} judge LLM. CompassJudger-1 is a general-purpose LLM that demonstrates remarkable versatility. It is capable of: 1. Performing unitary scoring and two-model comparisons as a reward model; 2. Conducting evaluations according to specified formats; 3. Generating critiques; 4. Executing diverse tasks like a general LLM. To assess the evaluation capabilities of different judge models under a unified setting, we have also established \textbf{JudgerBench}, a new benchmark that encompasses various subjective evaluation tasks and covers a wide range of topics. CompassJudger-1 offers a comprehensive solution for various evaluation tasks while maintaining the flexibility to adapt to diverse requirements. Both CompassJudger and JudgerBench are released and available to the research community athttps://github.com/open-compass/CompassJudger. We believe that by open-sourcing these tools, we can foster collaboration and accelerate progress in LLM evaluation methodologies.

cs.CL↗

arXiv2024

New constructions of DNA codes under multiple constraints and parallel searching algorithms

DNA codes have garnered significant interest due to their utilization in digital media storage, cryptography, and DNA computing. In this paper, we first extend the results of constructing reversible group codes \cite{Cengellenmis} and reversible composite group codes \cite{Korban5} to general even-order finite groups. By using these results, we give parallel searching algorithms to find some new DNA codes with better parameters. Secondly, by mapping codes over $\mathbb{F}_4$ to DNA codes, we establish a relationship between the $GC$-weight enumerator of DNA codes and the Hamming weight enumerator of their trace codes, which greatly improves the computational efficiency of searching for DNA codes. Based on this relationship, we propose an efficient algorithm for generating DNA codes with $50\%$ $GC$-content. Furthermore, we find that there is no direct connection between the $GC$-weight enumerator of a DNA code and the $GC$-weight enumerator of its dual code. Finally, we present algorithms for determining whether a DNA code is free from secondary structures or conflict-free, and some new DNA codes with better parameters under multiple constraints are obtained, which are listed in Tables 1 and 4.

cs.IT↗

arXiv2024

Optimal Few-GHW Linear Codes and Their Subcode Support Weight Distributions

Few-weight codes have been constructed and studied for many years, since their fascinating relations to finite geometries, strongly regular graphs and Boolean functions. Simplex codes are one-weight Griesmer $[\frac{q^k-1}{q-1},k ,q^{k-1}]_q$-linear codes and they meet all Griesmer bounds of the generalized Hamming weights of linear codes. All the subcodes with dimension $r$ of a $[\frac{q^k-1}{q-1},k ,q^{k-1}]_q$-simplex code have the same subcode support weight $\frac{q^{k-r}(q^r-1)}{q-1}$ for $1\leq r\leq k$. In this paper, we construct linear codes meeting the Griesmer bound of the $r$-generalized Hamming weight, such codes do not meet the Griesmer bound of the $j$-generalized Hamming weight for $1\leq j<r$. Moreover these codes have only few subcode support weights. The weight distribution and the subcode support weight distributions of these distance-optimal codes are determined. Linear codes constructed in this paper are natural generalizations of distance-optimal few-weight codes.

cs.IT↗

arXiv2024

Characterization of Intensity Correlation via Single-photon Detection in Quantum Key Distribution

One of the most significant vulnerabilities in the source unit of quantum key distribution (QKD) is the correlation between quantum states after modulation, which shall be characterized and evaluated for its practical security performance. In this work, we propose a methodology to characterize the intensity correlation according to the single-photon detection results in the measurement unit without modifying the configuration of the QKD system. In contrast to the previous research that employs extra classical optical detector to measure the correlation, our method can directly analyse the detection data generated during the raw key exchange, enabling to characterize the feature of correlation in real-time system operation. The basic method is applied to a BB84 QKD system and the characterized correlation decreases the secure key rate shown by the security proof. Furthermore, the method is extended and applied to characterize the correlation from the result of Bell-state measurement, which demonstrates its applicability to a running full-scheme MDI QKD system. This study provides an approach for standard certification of a QKD system.

quant-ph↗