Searcharxiv⌕ Search

arXiv subjects

Shiyang Zhang

Publications and source records attributed to Shiyang Zhang.

At least 19 recordsLinked to original sources
arXiv2026

Learning Interpretable Tumor Microenvironment Representations by Fitting Pan-Cancer Cell State-Niche Correlation

In the tumor microenvironment, cell's state is influenced by cell-cell interactions (CCIs) with neighboring cells in its niches. Identifying dysregulated CCIs that are associated with pathogenic process pinpoints targets for drug discovery. Imaging-based spatial transcriptomics and single-cell RNA sequencing provide, respectively, single-cell spatial information and transcriptome-wide measurements needed to study CCIs, but neither modality provides both. Existing spatial transcriptomics foundation models also cannot effectively learn from spatially resolved single-cell data with full-transcriptome coverage, explicitly infer the CCI mechanisms driving cell state-niche associations, or interpretable enough to support direct biological interpretations. Here, we present GITIII-scale, a hierarchical, interpretable pan-cancer spatial transcriptomics foundation model for TME representation learning that investigates cell state-niche associations and their underlying ligand-receptor (LR) signaling pathways. GITIII-scale uses transformers to model interactions between pairs of cells at defined spatial distances, an interpretable single-layer graph transformer without a feed-forward network to decompose how each gene in a receiver cell is influenced by each neighboring sender cell, and a graph transformer to generate cellular-neighborhood embeddings. Trained on our assembled pan-cancer database of specimen-matched scRNA-seq and imaging-based spatial transcriptomics datasets, GITIII-scale generated TME embeddings that recovered niche-associated state changes more accurately than existing spatial transcriptomics foundation models in cancer types unseen during training. A case study of an unseen breast cancer dataset further demonstrated the model's interpretability by identifying potentially drug-targetable LR pathways associated with endothelial overgrowth and tumorigenesis.

q-bio.GN↗
arXiv2026

MoRSE: Task-Oriented Multi-Agent System with Mixture of Role-Subtask Experts

Large language model-based multi-agent systems have recently shown strong potential for complex, long-horizon tasks. However, existing methods mainly rely on coarse prompt-level differentiation without parameter adaptation for diverse subtasks, resulting in insufficient inter-agent heterogeneity and limited specialized capability that bottleneck performance on tasks with complex requirements. To address this, we introduce a Task-Oriented Multi-Agent System with Mixture of Role-Subtask Experts (MoRSE) that distinguishes agents with (role, subtask)-conditional specialization at both the task structure and parameter levels. To make agents' responsibility explicit at the task structure level, we formulate a task-oriented multi-agent system that decomposes each task into a dependency-aware Directed Acyclic Graph of subtasks and assigns each agent a specific (role, subtask), introducing task-level specialization across collaborating agents. Additionally, to address the diverse role and subtask parameter adaptation demands, we propose a dynamic Mixture of (role, subtask) LoRA Experts module with a prototype-based semantic router for subtasks, augmenting agents with parameter-level specialization on a shared LLM substrate cost-effectively. Then, to co-optimize experts and router stably under sparse task rewards, we further propose a hierarchical group-relative policy optimization with two-layer credit assignment that isolates expert updates from the cross-route variance introduced by routing decisions, disentangling expert quality from routing quality. Experiments on code-generation benchmarks across three backbones demonstrate the effectiveness of our approach, with improvements in both whole-task and step-wise performance, and the gains from trained specialization generalize across held-out task categories and domains.

cs.MA↗
arXiv2026

On the Scaling of PEFT: Towards Million Personal Models of Trillion Parameters

Parameter-efficient fine-tuning (PEFT) is usually treated as a cheaper alternative to full fine-tuning. We study a broader role: small trainable adapters as persistent local state on top of strong shared foundation models. In this framing, the base model provides shared competence while adapters carry instance-specific behavior such as preferences, skills, tool habits, and memory-like updates. We organize the problem around three scaling axes: Scale Up, where stronger shared priors make small local updates more useful; Scale Down, where we study how small adapters can be while remaining reliable; and Scale Out, where many persistent adapted instances coexist. MinT provides one infrastructure example for managing adapter identity, revision, provenance, evaluation, and serving residency. Together, the results suggest that PEFT can be a compact substrate for persistent personal models rather than only a budget substitute for full fine-tuning.

cs.LG↗
arXiv2026

CoRE: A Fine-Grained Code Reasoning Benchmark Beyond Output Prediction

Despite strong performance on code generation tasks, it remains unclear whether large language models (LLMs) genuinely reason about code execution. Existing code reasoning benchmarks primarily evaluate final output correctness under a single canonical implementation, leaving two critical aspects underexplored: (1) whether LLMs can maintain consistency to functionally equivalent implementations, and (2) whether LLMs can accurately reason about intermediate execution states. We introduce \textbf{CoRE}, a \textbf{Co}de \textbf{Re}asoning benchmark that evaluates code reasoning through \textbf{implementation invariance} and \textbf{process transparency}. Extensive evaluations on eight frontier LLMs reveal two fundamental limitations. First, models exhibit a substantial \textbf{robustness gap}, with performance varying significantly across equivalent implementations. Second, we observe \textbf{superficial execution}, where models arrive at correct final outputs without correctly reasoning about intermediate execution states. Together, these findings demonstrate that output-only evaluations are insufficient for assessing code reasoning and position CoRE as a necessary benchmark for evaluating robust and faithful code reasoning.\footnote{Data and code are available at https://github.com/ZJUSig/CoRE.}

cs.SE↗
arXiv2026

Learning Permutation Distributions via Reflected Diffusion on Ranks

The finite symmetric group S_n provides a natural domain for permutations, yet learning probability distributions on S_n is challenging due to its factorially growing size and discrete, non-Euclidean structure. Recent permutation diffusion methods define forward noising via shuffle-based random walks (e.g., riffle shuffles) and learn reverse transitions with Plackett-Luce (PL) variants, but the resulting trajectories can be abrupt and increasingly hard to denoise as n grows. We propose Soft-Rank Diffusion, a discrete diffusion framework that replaces shuffle-based corruption with a structured soft-rank forward process: we lift permutations to a continuous latent representation of order by relaxing discrete ranks into soft ranks, yielding smoother and more tractable trajectories. For the reverse process, we introduce contextualized generalized Plackett-Luce (cGPL) denoisers that generalize prior PL-style parameterizations and improve expressivity for sequential decision structures. Experiments on sorting and combinatorial optimization benchmarks show that Soft-Rank Diffusion consistently outperforms prior diffusion baselines, with particularly strong gains in long-sequence and intrinsically sequential settings.

cs.LG↗
arXiv2026

STRIDE: Post-Training LLMs to Reason and Refine Bio-Sequences via Edit Trajectories

Discrete biological sequence optimization often requires goal-directed, parser-valid edits to an existing protein or molecule. Diffusion models support iterative refinement but do not expose a controllable discrete-edit interface, while autoregressive LLMs can be myopic when planning constrained edits over multiple steps. We introduce STRIDE (Sequence Trajectory Refinement via Iterative Discrete Editing), a post-training framework that trains an LLM to emit executable INSERT/DELETE/REPLACE trajectories for variable-length refinement. STRIDE first learns Levenshtein-aligned shortest-edit demonstrations, then uses supervised fine-tuning and group-based policy optimization to align trajectories with task rewards while preserving coherent editing. On an oracle-based full-action protein stress test, STRIDE raises success over Vanilla SFT from 42% to 89% and novelty among unique improvements from 47% to 97%. On instruction-conditioned molecular editing, the GSPO-aligned variant improves strict success, controllability, and SMILES validity over the SFT-only STRIDE model (code: https://github.com/daiheng-zhang/STRIDE).

cs.CE↗
arXiv2026

AutoQRA: Joint Optimization of Mixed-Precision Quantization and Low-rank Adapters for Efficient LLM Fine-Tuning

Quantization followed by parameter-efficient fine-tuning has emerged as a promising paradigm for downstream adaptation under tight GPU memory constraints. However, this sequential pipeline fails to leverage the intricate interaction between quantization bit-width and LoRA rank. Specifically, a carefully optimized quantization allocation with low quantization error does not always translate to strong fine-tuning performance, and different bit-width and rank configurations can lead to significantly varying outcomes under the same memory budget. To address this limitation, we propose AutoQRA, a joint optimization framework that simultaneously optimizes the bit-width and LoRA rank configuration for each layer during the mixed quantized fine-tuning process. To tackle the challenges posed by the large discrete search space and the high evaluation cost associated with frequent fine-tuning iterations, AutoQRA decomposes the optimization process into two stages. First, it first conducts a global multi-fidelity evolutionary search, where the initial population is warm-started by injecting layer-wise importance priors. This stage employs specific operators and a performance model to efficiently screen candidate configurations. Second, trust-region Bayesian optimization is applied to locally refine promising regions of the search space and identify optimal configurations under the given memory budget. This approach enables active compensation for quantization noise in specific layers during training. Experiments show that AutoQRA achieves performance close to full-precision fine-tuning with a memory footprint comparable to uniform 4-bit methods.

cs.LG↗
arXiv2026

Shared Nature, Unique Nurture: PRISM for Pluralistic Reasoning via In-context Structure Modeling

Large Language Models (LLMs) are converging towards a singular Artificial Hivemind, where shared Nature (pre-training priors) result in a profound collapse of distributional diversity, limiting the distinct perspectives necessary for creative exploration and scientific discovery. To address this, we propose to equip models with inference-time Nurture (individualized epistemic trajectories) using Epistemic Evolution paradigm, progressing through explore, internalize, and express. We instantiate this via PRISM (Pluralistic Reasoning via In-context Structure Modeling), a model-agnostic system that augments LLM with dynamic On-the-fly Epistemic Graphs. On three creativity benchmarks, PRISM achieves state-of-the-art novelty and significantly expands distributional diversity. Moreover, we evaluate the real-world utility via a challenging rare-disease diagnosis benchmark. Results demonstrate that PRISM successfully uncovers correct long-tail diagnoses that standard LLM miss, confirming that its divergence stems from meaningful exploration rather than incoherent noise. Overall, this work establishes a new paradigm for Pluralistic AI, moving beyond monolithic consensus toward a diverse ecosystem of unique cognitive individuals capable of collective, multi-perspective discovery.

cs.LG↗
arXiv2026

KiDS-Legacy: WIMP dark matter constraints from the cross-correlation of weak lensing and Fermi-LAT gamma rays

Dark matter dominates the matter content of the Universe, and its properties can be constrained through large-scale structure probes such as the cross-correlation between the unresolved gamma-ray background (UGRB) and weak gravitational lensing. We analysed 15 years of Fermi-LAT data, constructing UGRB intensity maps in ten energy bins (0.5-1000 GeV), and cross-correlated them with KiDS-Legacy shear in six tomographic bins. The measurements were performed using angular power spectra estimated with the pseudo-$C_\ell$ method. No significant cross-correlation is found. Based on this non-detection, we present 95% upper bounds on the weakly interacting massive particle (WIMP) decay rate $\Gamma_{\rm dec}$ and velocity-averaged annihilation cross-section $\langle\sigma_{\rm ann} v\rangle$ as functions of mass. We compare our results with bounds from other cosmological tracers and from local probes, and found them to be complementary, particularly at low masses ($\rm GeV/TeV$). In addition, using a Euclid-like lensing survey cross-correlated with Fermi-LAT, we forecast $\sim$2 times tighter limits, highlighting the potential of forthcoming data to strengthen constraints on dark matter annihilation and decay.

astro-ph.CO↗
arXiv2026

Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies

Deploying and fine-tuning Large Language Models (LLMs) on resource-constrained edge devices requires navigating a strict trade-off between memory footprint and task performance. While Quantization-Aware Fine-tuning has emerged as a viable solution, existing paradigms typically decouple quantization and adapter optimization. This separation overlooks a fundamental theoretical constraint we identify as the \textit{Fidelity-Plasticity Trade-off}: a layer's capacity to adapt to new tasks (Plasticity) is inherently constrained by the information capacity of its frozen weights (Fidelity). Aggressively quantizing semantically critical layers creates an information bottleneck that no amount of adapter rank can recover, while high precision in robust syntactic layers wastes valuable memory. To address this, we introduce \textbf{QR-Adaptor}, a unified framework that jointly optimizes per-layer quantization bit-width and LoRA rank. By formulating resource allocation as a multi-objective search aligned with the model's linguistic hierarchy, our method systematically liberates memory from redundancy-heavy layers to reinvest in capacity-critical ones. Extensive experiments demonstrate that QR-Adaptor establishes a new Pareto frontier: notably, a model fine-tuned under a strict 4-bit memory budget achieves performance rivaling 16-bit baselines, demonstrating that precise resource alignment is as critical as model size.

cs.LG↗
arXiv2025

Non-Markovian Discrete Diffusion with Causal Language Models

Discrete diffusion models offer a flexible, controllable approach to structured sequence generation, yet they still lag behind causal language models in expressive power. A key limitation lies in their reliance on the Markovian assumption, which restricts each step to condition only on the current state, leading to potential uncorrectable error accumulation. In this paper, we introduce CaDDi (Causal Discrete Diffusion Model), a discrete diffusion model that conditions on the entire generative trajectory, thereby lifting the Markov constraint and allowing the model to revisit and improve past states. By unifying sequential (causal) and temporal (diffusion) reasoning in a single non-Markovian transformer, CaDDi also treats standard causal language models as a special case and permits the direct reuse of pretrained LLM weights with no architectural changes. Empirically, CaDDi outperforms state-of-the-art discrete diffusion baselines on natural-language benchmarks, substantially narrowing the remaining gap to large autoregressive transformers.

cs.LG↗
arXiv2025

TANTE: Time-Adaptive Operator Learning via Neural Taylor Expansion

Operator learning for time-dependent partial differential equations (PDEs) has seen rapid progress in recent years, enabling efficient approximation of complex spatiotemporal dynamics. However, most existing methods rely on fixed time step sizes during rollout, which limits their ability to adapt to varying temporal complexity and often leads to error accumulation. Here, we propose the Time-Adaptive Transformer with Neural Taylor Expansion (TANTE), a novel operator-learning framework that produces continuous-time predictions with adaptive step sizes. TANTE predicts future states by performing a Taylor expansion at the current state, where neural networks learn both the higher-order temporal derivatives and the local radius of convergence. This allows the model to dynamically adjust its rollout based on the local behavior of the solution, thereby reducing cumulative error and improving computational efficiency. We demonstrate the effectiveness of TANTE across a wide range of PDE benchmarks, achieving superior accuracy and adaptability compared to fixed-step baselines, delivering accuracy gains of 60-80 % and speed-ups of 30-40 % at inference time. The code is publicly available at https://github.com/zwu88/TANTE for transparency and reproducibility.

cs.LG↗
arXiv2025

Blending effects on shear measurement synergy between Euclid-like and LSST-like surveys

Weak gravitational lensing is a powerful probe for constraining cosmological parameters, but its success relies on accurate shear measurements. In this paper, we use image simulations to investigate how a joint analysis of high-resolution space-based and deep ground-based imaging can improve shear estimation. We simulate two scenarios: a grid-based setup, where galaxies are placed on a regular grid to mimic an idealised, blending-free scenario, and a random setup, where galaxies are randomly distributed to capture the impact of blending. Comparing these cases, we find that blending introduces significant biases, particularly in LSST-like data due to its larger point spread function. This highlights the importance of including realistic blending effects when evaluating the performance of joint analyses. Using simulations that account for blending, we find that the most effective catalogue-level synergy is achieved by combining all galaxies detected in either survey. This approach yields an effective galaxy number density of $44.08~\rm arcmin^{-2}$ over the magnitude range of 20.0 to 27.5, compared to $39.17~\rm arcmin^{-2}$ for LSST-like data alone and $30.31~\rm arcmin^{-2}$ for \textit{Euclid}-like data alone. Restricting the analysis to only the overlapping sources detected in both surveys results in a gain of ${\sim}12\%$ compared to using either survey alone. While this joint-object approach is suboptimal at the catalogue level, it may become more effective in pixel-level analyses, where a joint fit to individual galaxy shapes can better leverage the complementary strengths of both data sets. Studying such pixel-level combinations, with realistic blending effects properly accounted for, remains a promising direction for future work.

astro-ph.CO↗
arXiv2025

Inaccuracy of an E-Dictionary and Its Influence on Chinese Language Users

Electronic dictionaries have largely replaced paper dictionaries and become central tools for L2 learners seeking to expand their vocabulary. Users often assume these resources are reliable and rarely question the validity of the definitions provided. The accuracy of major E-dictionaries is seldom scrutinized, and little attention has been paid to how their corpora are constructed. Research on dictionary use, particularly the limitations of electronic dictionaries, remains scarce. This study adopts a combined method of experimentation, user survey, and dictionary critique to examine Youdao, one of the most widely used E-dictionaries in China. The experiment involved a translation task paired with retrospective reflection. Participants were asked to translate sentences containing words that are insufficiently or inaccurately defined in Youdao. Their consultation behavior was recorded to analyze how faulty definitions influenced comprehension. Results show that incomplete or misleading definitions can cause serious misunderstandings. Additionally, students exhibited problematic consultation habits. The study further explores how such flawed definitions originate, highlighting issues in data processing and the integration of AI and machine learning technologies in dictionary construction. The findings suggest a need for better training in dictionary literacy for users, as well as improvements in the underlying AI models used to build E-dictionaries.

cs.CL↗
arXiv2025

Quantifying the Detectability of Milky Way Satellites with Image Simulations: a Case Study with KiDS

Ultra-faint dwarf galaxies, which can be detected as resolved satellite systems of the Milky Way, are critical to understanding galaxy formation, evolution, and the nature of dark matter, as they are the oldest, smallest, most metal-poor, and most dark matter-dominated stellar systems known. Quantifying the sensitivity of surveys is essential for understanding their capability and limitations in searching for ultra-faint satellites. In this paper, we present the first study of the image-level observational selection function for Kilo-Degree Survey (KiDS) based on the Synthetic UniveRses For Surveys (surfs)-based KiDS-Legacy-Like Simulations. We generate mock satellites and simulate images that include resolved stellar populations of the mock satellites and the background galaxies, capturing realistic observational effects such as source blending, photometric uncertainties, and star-galaxy separation. The matched-filter method is applied to recover the injected satellites. We derive the observational selection function of the survey in terms of the luminosity, half-light radius, and heliocentric distance of the satellites. Compared to a catalogue-level simulation as used in previous studies, the image-level simulation provides a more realistic assessment of survey sensitivity, accounting for observational limitations that are neglected in catalogue-level simulations. The image-level simulation shows a detection loss for compact sources with a distance $d \gtrsim 100~\rm kpc$. We argue that this is because compact sources are more likely to be identified as single sources rather than being resolved during the source extraction process.

astro-ph.CO↗
arXiv2025

Drop the Golden Apples: Identifying Third-Party Reuse by DB-Less Software Composition Analysis

The prevalent use of third-party libraries (TPLs) in modern software development introduces significant security and compliance risks, necessitating the implementation of Software Composition Analysis (SCA) to manage these threats. However, the accuracy of SCA tools heavily relies on the quality of the integrated feature database to cross-reference with user projects. While under the circumstance of the exponentially growing of open-source ecosystems and the integration of large models into software development, it becomes even more challenging to maintain a comprehensive feature database for potential TPLs. To this end, after referring to the evolution of LLM applications in terms of external data interactions, we propose the first framework of DB-Less SCA, to get rid of the traditional heavy database and embrace the flexibility of LLMs to mimic the manual analysis of security analysts to retrieve identical evidence and confirm the identity of TPLs by supportive information from the open Internet. Our experiments on two typical scenarios, native library identification for Android and copy-based TPL reuse for C/C++, especially on artifacts that are not that underappreciated, have demonstrated the favorable future for implementing database-less strategies in SCA.

cs.SE↗
arXiv2025

Intelligence at the Edge of Chaos

We explore the emergence of intelligent behavior in artificial systems by investigating how the complexity of rule-based systems influences the capabilities of models trained to predict these rules. Our study focuses on elementary cellular automata (ECA), simple yet powerful one-dimensional systems that generate behaviors ranging from trivial to highly complex. By training distinct Large Language Models (LLMs) on different ECAs, we evaluated the relationship between the complexity of the rules' behavior and the intelligence exhibited by the LLMs, as reflected in their performance on downstream tasks. Our findings reveal that rules with higher complexity lead to models exhibiting greater intelligence, as demonstrated by their performance on reasoning and chess move prediction tasks. Both uniform and periodic systems, and often also highly chaotic systems, resulted in poorer downstream performance, highlighting a sweet spot of complexity conducive to intelligence. We conjecture that intelligence arises from the ability to predict complexity and that creating intelligence may require only exposure to complexity.

cs.AI↗
arXiv2024

AutoMixQ: Self-Adjusting Quantization for High Performance Memory-Efficient Fine-Tuning

Fine-tuning large language models (LLMs) under resource constraints is a significant challenge in deep learning. Low-Rank Adaptation (LoRA), pruning, and quantization are all effective methods for improving resource efficiency. However, combining them directly often results in suboptimal performance, especially with uniform quantization across all model layers. This is due to the complex, uneven interlayer relationships introduced by pruning, necessitating more refined quantization strategies. To address this, we propose AutoMixQ, an end-to-end optimization framework that selects optimal quantization configurations for each LLM layer. AutoMixQ leverages lightweight performance models to guide the selection process, significantly reducing time and computational resources compared to exhaustive search methods. By incorporating Pareto optimality, AutoMixQ balances memory usage and performance, approaching the upper bounds of model capability under strict resource constraints. Our experiments on widely used benchmarks show that AutoMixQ reduces memory consumption while achieving superior performance. For example, at a 30\% pruning rate in LLaMA-7B, AutoMixQ achieved 66.21\% on BoolQ compared to 62.45\% for LoRA and 58.96\% for LoftQ, while reducing memory consumption by 35.5\% compared to LoRA and 27.5\% compared to LoftQ.

cs.LG↗