arXiv subjects

Haoyang Liu

Publications and source records attributed to Haoyang Liu.

At least 55 records · Page 3Linked to original sources

arXiv2025

Toward a Team of AI-made Scientists for Scientific Discovery from Gene Expression Data

Machine learning has emerged as a powerful tool for scientific discovery, enabling researchers to extract meaningful insights from complex datasets. For instance, it has facilitated the identification of disease-predictive genes from gene expression data, significantly advancing healthcare. However, the traditional process for analyzing such datasets demands substantial human effort and expertise for the data selection, processing, and analysis. To address this challenge, we introduce a novel framework, a Team of AI-made Scientists (TAIS), designed to streamline the scientific discovery pipeline. TAIS comprises simulated roles, including a project manager, data engineer, and domain expert, each represented by a Large Language Model (LLM). These roles collaborate to replicate the tasks typically performed by data scientists, with a specific focus on identifying disease-predictive genes. Furthermore, we have curated a benchmark dataset to assess TAIS's effectiveness in gene identification, demonstrating our system's potential to significantly enhance the efficiency and scope of scientific exploration. Our findings represent a solid step towards automating scientific discovery through large language models.

q-bio.GN↗

arXiv2025

Rethinking Selectivity in State Space Models: A Minimal Predictive Sufficiency Approach

State Space Models (SSMs), particularly recent selective variants like Mamba, have emerged as a leading architecture for sequence modeling, challenging the dominance of Transformers. However, the success of these state-of-the-art models largely relies on heuristically designed selective mechanisms, which lack a rigorous first-principle derivation. This theoretical gap raises questions about their optimality and robustness against spurious correlations. To address this, we introduce the Principle of Predictive Sufficiency, a novel information-theoretic criterion stipulating that an ideal hidden state should be a minimal sufficient statistic of the past for predicting the future. Based on this principle, we propose the Minimal Predictive Sufficiency State Space Model (MPS-SSM), a new framework where the selective mechanism is guided by optimizing an objective function derived from our principle. This approach encourages the model to maximally compress historical information without losing predictive power, thereby learning to ignore non-causal noise and spurious patterns. Extensive experiments on a wide range of benchmark datasets demonstrate that MPS-SSM not only achieves state-of-the-art performance, significantly outperforming existing models in long-term forecasting and noisy scenarios, but also exhibits superior robustness. Furthermore, we show that the MPS principle can be extended as a general regularization framework to enhance other popular architectures, highlighting its broad potential.

cs.LG↗

arXiv2025

X-Intelligence 3.0: Training and Evaluating Reasoning LLM for Semiconductor Display

Large language models (LLMs) have recently achieved significant advances in reasoning and demonstrated their advantages in solving challenging problems. Yet, their effectiveness in the semiconductor display industry remains limited due to a lack of domain-specific training and expertise. To bridge this gap, we present X-Intelligence 3.0, the first high-performance reasoning model specifically developed for the semiconductor display industry. This model is designed to deliver expert-level understanding and reasoning for the industry's complex challenges. Leveraging a carefully curated industry knowledge base, the model undergoes supervised fine-tuning and reinforcement learning to enhance its reasoning and comprehension capabilities. To further accelerate development, we implemented an automated evaluation framework that simulates expert-level assessments. We also integrated a domain-specific retrieval-augmented generation (RAG) mechanism, resulting in notable performance gains on benchmark datasets. Despite its relatively compact size of 32 billion parameters, X-Intelligence 3.0 outperforms SOTA DeepSeek-R1-671B across multiple evaluations. This demonstrates its exceptional efficiency and establishes it as a powerful solution to the longstanding reasoning challenges faced by the semiconductor display industry.

cs.CL↗

arXiv2025

Splitting of abelian varieties in motivic stable homotopy category

In this paper, we discuss the motivic stable homotopy type of abelian varieties. For an abelian variety over a perfect field $k$ with a rational point, it always splits off a top-dimensional cell in motivic stable homotopy category $\text{SH}(k)$. Let $k=\mathbb{R}$, there is a concrete splitting which is determined by the motive of X and the real points $X(\mathbb{R})$ in $\text{SH}(\mathbb{R})_Λ$ for some $\mathbb{Z}\subsetΛ\subset\mathbb{Q}$. We will also discuss this splitting from a viewpoint of the Chow-Witt correspondences.

math.AG↗

arXiv2025

On the splitting of surfaces in motivic stable homotopy category

Let $k$ be a perfect field and $X$ be a smooth projective surface over $k$ with a rational point, we discuss the condition of splitting off the top cell for the motivic stable homotopy type of $X$. We also study some outlying examples, such as K3 surfaces. When $k$ is an algebraically field with characteristic not equal to 2, we can give an alternative proof of the splitting result of curves and also understand the splittings of Calabi-Yau surfaces via the motivic Hurewicz theorem and decomposition of the Chow-Witt correspondences.

math.AG↗

arXiv2025

Dataset Distillation via the Wasserstein Metric

Dataset Distillation (DD) aims to generate a compact synthetic dataset that enables models to achieve performance comparable to training on the full large dataset, significantly reducing computational costs. Drawing from optimal transport theory, we introduce WMDD (Wasserstein Metric-based Dataset Distillation), a straightforward yet powerful method that employs the Wasserstein metric to enhance distribution matching. We compute the Wasserstein barycenter of features from a pretrained classifier to capture essential characteristics of the original data distribution. By optimizing synthetic data to align with this barycenter in feature space and leveraging per-class BatchNorm statistics to preserve intra-class variations, WMDD maintains the efficiency of distribution matching approaches while achieving state-of-the-art results across various high-resolution datasets. Our extensive experiments demonstrate WMDD's effectiveness and adaptability, highlighting its potential for advancing machine learning applications at scale.

cs.CV↗

arXiv2025

CNN-Derived Elemental Abundances of LAMOST DR10 Giants: Implications for Galactic Substructures

Stellar parameters and abundances provide crucial insights into stellar and Galactic evolution studies. In this work, we developed a convolutional neural network (CNN) to estimate stellar parameters: effective temperature ($T_{\text{eff}}$), surface gravity (log $g$) and metallicity (both [Fe/H] and [M/H]) as well as six $α$-elements (C, N, O, Mg, Si, Ca) and [$α$/M]. We selected giant stars with $ 3500 \, \text{K} < T_{\text{eff}} < 5500 \, \text{K} $ and $ 0 \, \text{dex} < \log g < 3.6 \, \text{dex} $ from the LAMOST and APOGEE surveys, while requiring (S/N)$_g$ of the LAMOST low-resolution spectra $>$ 10, which leaves 1,100,858 giant stars. The spectral from LAMOST and the labels from APOGEE for 62,511 common stars were used as our training set. The corresponding test set yields scatters 50 K, 0.06 dex and 0.13 dex for $T_{\text{eff}}$, [Fe/H] and log $g$, respectively. For $α$ elements O, Mg, Si and Ca, the scatters are 0.05 dex, 0.04 dex, 0.03 and 0.04 dex, respectively. For C and N elements, the scatters are 0.07 dex and 0.05 dex. For [$α$/M] and [M/H], the scatters are 0.03 dex and 0.06 dex. The mean absolute error (MAE) of most elements are between 0.02 $-$ 0.04 dex. The predicted abundances were cross-matched with previously identified substructures PG1 and PG2, with their origins subsequently analyzed. Finally, the catalog is available at https://nadc.china-vo.org/res/r101529/.

astro-ph.GA↗

arXiv2025

Cellular $\mathbb{A}^1$-Homology of Smooth Toric Varieties

In this paper, we present the calculations of cellular $\mathbb{A}^1$-homology for smooth toric varieties, along with an explicit description of pure shellable cases. Consequently, we derive the (Milnor-Witt) motivic decomposition for these pure shellable cases. Furthermore, we obtain an additive basis for the Chow groups of general smooth toric varieties.

math.AG↗

arXiv2025

GenoTEX: An LLM Agent Benchmark for Automated Gene Expression Data Analysis

Recent advancements in machine learning have significantly improved the identification of disease-associated genes from gene expression datasets. However, these processes often require extensive expertise and manual effort, limiting their scalability. Large Language Model (LLM)-based agents have shown promise in automating these tasks due to their increasing problem-solving abilities. To support the evaluation and development of such methods, we introduce GenoTEX, a benchmark dataset for the automated analysis of gene expression data. GenoTEX provides analysis code and results for solving a wide range of gene-trait association problems, encompassing dataset selection, preprocessing, and statistical analysis, in a pipeline that follows computational genomics standards. The benchmark includes expert-curated annotations from bioinformaticians to ensure accuracy and reliability. To provide baselines for these tasks, we present GenoAgent, a team of LLM-based agents that adopt a multi-step programming workflow with flexible self-correction, to collaboratively analyze gene expression datasets. Our experiments demonstrate the potential of LLM-based methods in analyzing genomic data, while error analysis highlights the challenges and areas for future improvement. We propose GenoTEX as a promising resource for benchmarking and enhancing automated methods for gene expression data analysis. The benchmark is available at https://github.com/Liu-Hy/GenoTEX.

cs.LG↗

arXiv2025

Towards Adversarially Robust Dataset Distillation by Curvature Regularization

Dataset distillation (DD) allows datasets to be distilled to fractions of their original size while preserving the rich distributional information, so that models trained on the distilled datasets can achieve a comparable accuracy while saving significant computational loads. Recent research in this area has been focusing on improving the accuracy of models trained on distilled datasets. In this paper, we aim to explore a new perspective of DD. We study how to embed adversarial robustness in distilled datasets, so that models trained on these datasets maintain the high accuracy and meanwhile acquire better adversarial robustness. We propose a new method that achieves this goal by incorporating curvature regularization into the distillation process with much less computational overhead than standard adversarial training. Extensive empirical experiments suggest that our method not only outperforms standard adversarial training on both accuracy and robustness with less computation overhead but is also capable of generating robust distilled datasets that can withstand various adversarial attacks. Our implementation is available at: https://github.com/yumozi/GUARD.

cs.LG↗

arXiv2025

Chemical Components in the Virgo Overdensity and Hercules-Aquila Cloud: hints of more than one merger event in GSE-like debris

Using elemental abundances for 1.26 million K giants in the LAMOST DR8 value-added catalog, we analyze the chemical abundances of the Virgo Overdensity (VOD) and Hercules-Aquila Cloud (HAC). We find two distinct chemical populations in both overdensities, which is in disagreement with the mainstream hypothesis that both overdensities are composed of materials from a single merger event, namely Gaia-Sausage-Enceladus (GSE). The two populations show different chemical trends: one exhibits low metallicities and high $α$ abundances, and the other shows high metallicities and low $α$ abundances, which is associated with the recently discovered Nereus and Virgo Radial Merger (VRM) components in the local stellar halo, respectively. The Nereus component in these overdensities uniquely exhibits a decreasing trend in the [Fe/H]-[Mn/Fe] plane. Out of all observed Milky Way dwarf galaxies, this trend is only found in the Sculptor dwarf galaxy, which provides clues for the properties of Nereus progenitor. We also find that the velocity ellipse with high aniostropy parameters that is usually considered to be part of GSE are actually a mix of the two components. Both overdensities are well-mixed in kinematic spaces, confirming recent claims that the debris of merger pairs are kinematically indistinguishable in a recent simulation. We find that the velocity ellipses of the VRM stars in these overdensities have large inclination angles, which may be an indication of the merger time in simulated merger events.

astro-ph.GA↗

arXiv2025

Investigate Substructures of the Inner Halo Using the StarGO Algorithm with APOGEE DR17

We applied Stars' Galactic Origin (StarGO) algorithm in 7D space (i.e., [Fe/H], [Mg/Fe], [Al/Fe], $L_{z}$, $J_{r}$, $J_{z}$, $E$) to analyze stars in the inner halo with APOGEE DR17. We identified some known substructures, including Gaia-Sausage-Enceladus (GSE), Sagittarius Stream (Sgr), LMS-1 (Wukong), Thamnos, Metal-Weak Thick Disk (MWTD) and Aleph. Additionally, we identified an undefined metal-poor group (UDG, with [Fe/H] $<-0.8$ dex) probably linked to known substructure like Aleph, as well as a high $α$-abundance substructure (HAS) associated with both the Nyx and Nyx-2 streams. Chemical abundance of the HAS supports the argument that Nyx and Nyx-2 share a common origin. We discovered three substructures, which we refer to as new-substructure candidate-1,2,3 (NSTC-1,2,3). Despite exhibiting disk-like dynamics, these NSTCs demonstrate notably low [Mg/Fe] ($<$ 0.2 dex) and [Al/Fe] ($<$ $-0.15$ dex), similar to the properties of dwarf galaxies. Their high orbital energy and low [$α$/Fe] indicate the association with recent accretion events.

astro-ph.GA↗

arXiv2025

Approximate Nullspace Augmented Finetuning for Robust Vision Transformers

Enhancing the robustness of deep learning models, particularly in the realm of vision transformers (ViTs), is crucial for their real-world deployment. In this work, we provide a finetuning approach to enhance the robustness of vision transformers inspired by the concept of nullspace from linear algebra. Our investigation centers on whether a vision transformer can exhibit resilience to input variations akin to the nullspace property in linear mappings, which would imply that perturbations sampled from this nullspace do not influence the model's output when added to the input. We start from the observation that many existing ViTs satisfy this property because their patch embedding layer has a non-trivial nullspace. Then, we extend the notion of nullspace to nonlinear settings and demonstrate that it is possible to synthesize approximate nullspace elements for ViT's encoder blocks through optimization. Finally, we propose a finetuning strategy for ViTs wherein we augment the training data with synthesized approximate nullspace noise. We find that our finetuning approach significantly improves the models' robustness to both adversarial and natural image perturbations.\footnote{Code is available at: https://github.com/Liu-Hy/ns-vit.

cs.CV↗

arXiv2025

Apollo-MILP: An Alternating Prediction-Correction Neural Solving Framework for Mixed-Integer Linear Programming

Leveraging machine learning (ML) to predict an initial solution for mixed-integer linear programming (MILP) has gained considerable popularity in recent years. These methods predict a solution and fix a subset of variables to reduce the problem dimension. Then, they solve the reduced problem to obtain the final solutions. However, directly fixing variable values can lead to low-quality solutions or even infeasible reduced problems if the predicted solution is not accurate enough. To address this challenge, we propose an Alternating prediction-correction neural solving framework (Apollo-MILP) that can identify and select accurate and reliable predicted values to fix. In each iteration, Apollo-MILP conducts a prediction step for the unfixed variables, followed by a correction step to obtain an improved solution (called reference solution) through a trust-region search. By incorporating the predicted and reference solutions, we introduce a novel Uncertainty-based Error upper BOund (UEBO) to evaluate the uncertainty of the predicted values and fix those with high confidence. A notable feature of Apollo-MILP is the superior ability for problem reduction while preserving optimality, leading to high-quality final solutions. Experiments on commonly used benchmarks demonstrate that our proposed Apollo-MILP significantly outperforms other ML-based approaches in terms of solution quality, achieving over a 50% reduction in the solution gap.

cs.LG↗

arXiv2025

Signatures of Chiral Superconductivity in Rhombohedral Graphene

Chiral superconductors are unconventional superconducting states that break time reversal symmetry spontaneously and typically feature Cooper pairing at non-zero angular momentum. Such states may host Majorana fermions and provide an important platform for topological physics research and fault-tolerant quantum computing. Despite intensive search and prolonged studies of several candidate systems, chiral superconductivity has remained elusive so far. Here we report the discovery of robust unconventional superconductivity in rhombohedral tetra- and penta-layer graphene in the absence of moiré superlattice effects. We observed two superconducting states in the gate-induced flat conduction bands with Tc up to 300 mK and charge density ne as low as 2.4*1011 cm-2 in three tetralayer and two pentalayer devices. Spontaneous time-reversal-symmetry-breaking (TRSB) due to electron's orbital motion is found, and several observations indicate the chiral nature of these superconducting states, including: 1. In the superconducting state, Rxx shows magnetic hysteresis in varying out-of-plane magnetic field B, which is absent from all other superconductors; 2. the superconducting states are immune to in-plane magnetic field and are developed within a spin- and valley-polarized quarter-metal phase; 3. the normal states show anomalous Hall signals at zero magnetic field and magnetic hysteresis. We also observed a critical B of up to 1.4 Tesla, higher than any graphene superconductivity reported so far and indicates a strong-coupling superconductivity close to the BCS-BEC crossover. Our observations establish a pure carbon material for the study of topological superconductivity, and pave the way to explore Majorana modes and topological quantum computing.

cond-mat.mes-hall↗

arXiv2025

Invariant Graph Learning Meets Information Bottleneck for Out-of-Distribution Generalization

Graph out-of-distribution (OOD) generalization remains a major challenge in graph learning since graph neural networks (GNNs) often suffer from severe performance degradation under distribution shifts. Invariant learning, aiming to extract invariant features across varied distributions, has recently emerged as a promising approach for OOD generation. Despite the great success of invariant learning in OOD problems for Euclidean data (i.e., images), the exploration within graph data remains constrained by the complex nature of graphs. Existing studies, such as data augmentation or causal intervention, either suffer from disruptions to invariance during the graph manipulation process or face reliability issues due to a lack of supervised signals for causal parts. In this work, we propose a novel framework, called Invariant Graph Learning based on Information bottleneck theory (InfoIGL), to extract the invariant features of graphs and enhance models' generalization ability to unseen distributions. Specifically, InfoIGL introduces a redundancy filter to compress task-irrelevant information related to environmental factors. Cooperating with our designed multi-level contrastive learning, we maximize the mutual information among graphs of the same class in the downstream classification tasks, preserving invariant features for prediction to a great extent. An appealing feature of InfoIGL is its strong generalization ability without depending on supervised signal of invariance. Experiments on both synthetic and real-world datasets demonstrate that our method achieves state-of-the-art performance under OOD generalization for graph classification tasks. The source code is available at https://github.com/maowenyu-11/InfoIGL.

cs.LG↗

arXiv2025

Distinguished Quantized Guidance for Diffusion-based Sequence Recommendation

Diffusion models (DMs) have emerged as promising approaches for sequential recommendation due to their strong ability to model data distributions and generate high-quality items. Existing work typically adds noise to the next item and progressively denoises it guided by the user's interaction sequence, generating items that closely align with user interests. However, we identify two key issues in this paradigm. First, the sequences are often heterogeneous in length and content, exhibiting noise due to stochastic user behaviors. Using such sequences as guidance may hinder DMs from accurately understanding user interests. Second, DMs are prone to data bias and tend to generate only the popular items that dominate the training dataset, thus failing to meet the personalized needs of different users. To address these issues, we propose Distinguished Quantized Guidance for Diffusion-based Sequence Recommendation (DiQDiff), which aims to extract robust guidance to understand user interests and generate distinguished items for personalized user interests within DMs. To extract robust guidance, DiQDiff introduces Semantic Vector Quantization (SVQ) to quantize sequences into semantic vectors (e.g., collaborative signals and category interests) using a codebook, which can enrich the guidance to better understand user interests. To generate distinguished items, DiQDiff personalizes the generation through Contrastive Discrepancy Maximization (CDM), which maximizes the distance between denoising trajectories using contrastive loss to prevent biased generation for different users. Extensive experiments are conducted to compare DiQDiff with multiple baseline models across four widely-used datasets. The superior recommendation performance of DiQDiff against leading approaches demonstrates its effectiveness in sequential recommendation tasks.

cs.IR↗

arXiv2024

MILP-StuDio: MILP Instance Generation via Block Structure Decomposition

Mixed-integer linear programming (MILP) is one of the most popular mathematical formulations with numerous applications. In practice, improving the performance of MILP solvers often requires a large amount of high-quality data, which can be challenging to collect. Researchers thus turn to generation techniques to generate additional MILP instances. However, existing approaches do not take into account specific block structures -- which are closely related to the problem formulations -- in the constraint coefficient matrices (CCMs) of MILPs. Consequently, they are prone to generate computationally trivial or infeasible instances due to the disruptions of block structures and thus problem formulations. To address this challenge, we propose a novel MILP generation framework, called Block Structure Decomposition (MILP-StuDio), to generate high-quality instances by preserving the block structures. Specifically, MILP-StuDio begins by identifying the blocks in CCMs and decomposing the instances into block units, which serve as the building blocks of MILP instances. We then design three operators to construct new instances by removing, substituting, and appending block units in the original instances, enabling us to generate instances with flexible sizes. An appealing feature of MILP-StuDio is its strong ability to preserve the feasibility and computational hardness of the generated instances. Experiments on the commonly-used benchmarks demonstrate that using instances generated by MILP-StuDio is able to significantly reduce over 10% of the solving time for learning-based solvers.

cs.LG↗