arXiv subjects

Jiaqi Wei

Publications and source records attributed to Jiaqi Wei.

35 records · Page 2Linked to original sources

arXiv2025

Bidirectional Representations Augmented Autoregressive Biological Sequence Generation

Autoregressive (AR) models, common in sequence generation, are limited in many biological tasks such as de novo peptide sequencing and protein modeling by their unidirectional nature, failing to capture crucial global bidirectional token dependencies. Non-Autoregressive (NAR) models offer holistic, bidirectional representations but face challenges with generative coherence and scalability. To transcend this, we propose a hybrid framework enhancing AR generation by dynamically integrating rich contextual information from non-autoregressive mechanisms. Our approach couples a shared input encoder with two decoders: a non-autoregressive one learning latent bidirectional biological features, and an AR decoder synthesizing the biological sequence by leveraging these bidirectional features. A novel cross-decoder attention module enables the AR decoder to iteratively query and integrate these bidirectional features, enriching its predictions. This synergy is cultivated via a tailored training strategy with importance annealing for balanced objectives and cross-decoder gradient blocking for stable, focused learning. Evaluations on a demanding nine-species benchmark of de novo peptide sequencing show that our model substantially surpasses AR and NAR baselines. It uniquely harmonizes AR stability with NAR contextual awareness, delivering robust, superior performance on diverse downstream data. This research advances biological sequence modeling techniques and contributes a novel architectural paradigm for augmenting AR models with enhanced bidirectional understanding for complex sequence generation. Code is available at https://github.com/BEAM-Labs/denovo.

cs.LG↗

arXiv2025

Scientists' First Exam: Probing Cognitive Abilities of MLLM via Perception, Understanding, and Reasoning

Scientific discoveries increasingly rely on complex multimodal reasoning based on information-intensive scientific data and domain-specific expertise. Empowered by expert-level scientific benchmarks, scientific Multimodal Large Language Models (MLLMs) hold the potential to significantly enhance this discovery process in realistic workflows. However, current scientific benchmarks mostly focus on evaluating the knowledge understanding capabilities of MLLMs, leading to an inadequate assessment of their perception and reasoning abilities. To address this gap, we present the Scientists' First Exam (SFE) benchmark, designed to evaluate the scientific cognitive capacities of MLLMs through three interconnected levels: scientific signal perception, scientific attribute understanding, scientific comparative reasoning. Specifically, SFE comprises 830 expert-verified VQA pairs across three question types, spanning 66 multimodal tasks across five high-value disciplines. Extensive experiments reveal that current state-of-the-art GPT-o3 and InternVL-3 achieve only 34.08% and 26.52% on SFE, highlighting significant room for MLLMs to improve in scientific realms. We hope the insights obtained in SFE will facilitate further developments in AI-enhanced scientific discoveries.

cs.AI↗

arXiv2025

Single femtosecond laser pulse-driven ferromagnetic switching

Light pulses offer a faster, more energy-efficient, and direct route to magnetic bit writing, pointing toward a hybrid memory and computing paradigm based on photon transmission and spin retention. Yet progress remains hindered, as deterministic, single-pulse optical toggle switching has so far been achieved only with ferrimagnetic materials, which require too specific a rare-earth composition and temperature conditions for technological use. In mainstream ferromagnet--central to spintronic memory and storage--such bistable switching is considered fundamentally difficult, as laser-induced heating does not inherently break time-reversal symmetry. Here, we report coherent magnetization switching in ferromagnets, driven by thermal anisotropy torque with single laser pulses. The toggle switching behavior is robust over a broad range of pulse durations, from femtoseconds to picoseconds, a prerequisite for practical applications. Furthermore, the phenomenon exhibits reproducibility in CoFeB/MgO-based magnetic tunnel junctions with a high magnetoresistance exceeding 110%, as well as the scalability down to nanoscales with remarkable energy efficiency (17 fJ per 100-nm-sized bit). These results mark a notable step toward integrating opto-spintronics into next-generation memory and storage technologies.

cond-mat.mes-hall↗

arXiv2025

From AI for Science to Agentic Science: A Survey on Autonomous Scientific Discovery

Artificial intelligence (AI) is reshaping scientific discovery, evolving from specialized computational tools into autonomous research partners. We position Agentic Science as a pivotal stage within the broader AI for Science paradigm, where AI systems progress from partial assistance to full scientific agency. Enabled by large language models (LLMs), multimodal systems, and integrated research platforms, agentic AI shows capabilities in hypothesis generation, experimental design, execution, analysis, and iterative refinement -- behaviors once regarded as uniquely human. This survey provides a domain-oriented review of autonomous scientific discovery across life sciences, chemistry, materials science, and physics. We unify three previously fragmented perspectives -- process-oriented, autonomy-oriented, and mechanism-oriented -- through a comprehensive framework that connects foundational capabilities, core processes, and domain-specific realizations. Building on this framework, we (i) trace the evolution of AI for Science, (ii) identify five core capabilities underpinning scientific agency, (iii) model discovery as a dynamic four-stage workflow, (iv) review applications across the above domains, and (v) synthesize key challenges and future opportunities. This work establishes a domain-oriented synthesis of autonomous scientific discovery and positions Agentic Science as a structured paradigm for advancing AI-driven research.

cs.LG↗

arXiv2025

A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers

Scientific Large Language Models (Sci-LLMs) are transforming how knowledge is represented, integrated, and applied in scientific research, yet their progress is shaped by the complex nature of scientific data. This survey presents a comprehensive, data-centric synthesis that reframes the development of Sci-LLMs as a co-evolution between models and their underlying data substrate. We formulate a unified taxonomy of scientific data and a hierarchical model of scientific knowledge, emphasizing the multimodal, cross-scale, and domain-specific challenges that differentiate scientific corpora from general natural language processing datasets. We systematically review recent Sci-LLMs, from general-purpose foundations to specialized models across diverse scientific disciplines, alongside an extensive analysis of over 270 pre-/post-training datasets, showing why Sci-LLMs pose distinct demands -- heterogeneous, multi-scale, uncertainty-laden corpora that require representations preserving domain invariance and enabling cross-modal reasoning. On evaluation, we examine over 190 benchmark datasets and trace a shift from static exams toward process- and discovery-oriented assessments with advanced evaluation protocols. These data-centric analyses highlight persistent issues in scientific data development and discuss emerging solutions involving semi-automated annotation pipelines and expert validation. Finally, we outline a paradigm shift toward closed-loop systems where autonomous agents based on Sci-LLMs actively experiment, validate, and contribute to a living, evolving knowledge base. Collectively, this work provides a roadmap for building trustworthy, continually evolving artificial intelligence (AI) systems that function as a true partner in accelerating scientific discovery.

cs.CL↗

arXiv2025

Retrieval is Not Enough: Enhancing RAG Reasoning through Test-Time Critique and Optimization

Retrieval-augmented generation (RAG) has become a widely adopted paradigm for enabling knowledge-grounded large language models (LLMs). However, standard RAG pipelines often fail to ensure that model reasoning remains consistent with the evidence retrieved, leading to factual inconsistencies or unsupported conclusions. In this work, we reinterpret RAG as Retrieval-Augmented Reasoning and identify a central but underexplored problem: Reasoning Misalignment -- the divergence between an LLM's internal reasoning trajectory and the evidential constraints provided by retrieval. To address this issue, we propose AlignRAG, a novel iterative framework grounded in Critique-Driven Alignment (CDA). We further introduce AlignRAG-auto, an autonomous variant that dynamically terminates refinement, removing the need to pre-specify the number of critique iterations. At the heart of AlignRAG lies a contrastive critique synthesis mechanism that generates retrieval-sensitive critiques while mitigating self-bias. This mechanism trains a dedicated retrieval-augmented Critic Language Model (CLM) using labeled critiques that distinguish between evidence-aligned and misaligned reasoning. Empirical evaluations show that our approach significantly improves reasoning fidelity. Our 8B-parameter CLM improves performance over the Self-Refine baseline by 12.1% on out-of-domain tasks and outperforms a standard 72B-parameter CLM by 2.2%. Furthermore, AlignRAG-auto achieves this state-of-the-art performance while dynamically determining the optimal number of refinement steps, enhancing efficiency and usability. AlignRAG remains compatible with existing RAG architectures as a plug-and-play module and demonstrates strong robustness under both informative and noisy retrieval scenarios.

cs.AI↗

arXiv2025

Unifying Tree Search Algorithm and Reward Design for LLM Reasoning: A Survey

Deliberative tree search is a cornerstone of modern Large Language Model (LLM) research, driving the pivot from brute-force scaling toward algorithmic efficiency. This single paradigm unifies two critical frontiers: \textbf{Test-Time Scaling (TTS)}, which deploys on-demand computation to solve hard problems, and \textbf{Self-Improvement}, which uses search-generated data to durably enhance model parameters. However, this burgeoning field is fragmented and lacks a common formalism, particularly concerning the ambiguous role of the reward signal -- is it a transient heuristic or a durable learning target? This paper resolves this ambiguity by introducing a unified framework that deconstructs search algorithms into three core components: the \emph{Search Mechanism}, \emph{Reward Formulation}, and \emph{Transition Function}. We establish a formal distinction between transient \textbf{Search Guidance} for TTS and durable \textbf{Parametric Reward Modeling} for Self-Improvement. Building on this formalism, we introduce a component-centric taxonomy, synthesize the state-of-the-art, and chart a research roadmap toward more systematic progress in creating autonomous, self-improving agents.

cs.CL↗

arXiv2025

TimeSeriesScientist: A General-Purpose AI Agent for Time Series Analysis

Time series forecasting is central to decision-making in domains as diverse as energy, finance, climate, and public health. In practice, forecasters face thousands of short, noisy series that vary in frequency, quality, and horizon, where the dominant cost lies not in model fitting, but in the labor-intensive preprocessing, validation, and ensembling required to obtain reliable predictions. Prevailing statistical and deep learning models are tailored to specific datasets or domains and generalize poorly. A general, domain-agnostic framework that minimizes human intervention is urgently in demand. In this paper, we introduce TimeSeriesScientist (TSci), the first LLM-driven agentic framework for general time series forecasting. The framework comprises four specialized agents: Curator performs LLM-guided diagnostics augmented by external tools that reason over data statistics to choose targeted preprocessing; Planner narrows the hypothesis space of model choice by leveraging multi-modal diagnostics and self-planning over the input; Forecaster performs model fitting and validation and, based on the results, adaptively selects the best model configuration as well as ensemble strategy to make final predictions; and Reporter synthesizes the whole process into a comprehensive, transparent report. With transparent natural-language rationales and comprehensive reports, TSci transforms the forecasting workflow into a white-box system that is both interpretable and extensible across tasks. Empirical results on eight established benchmarks demonstrate that TSci consistently outperforms both statistical and LLM-based baselines, reducing forecast error by an average of 10.4% and 38.2%, respectively. Moreover, TSci produces a clear and rigorous report that makes the forecasting workflow more transparent and interpretable.

cs.LG↗

arXiv2025

Single-shot optical precessional magnetization switching of Pt/Co/Pt ferromagnetic trilayers

Ultra-fast magnetization switching triggered by a single femtosecond laser pulse has gained significant attention over the last decade for its potential in low-power consumption, high-speed memory applications. However, this phenomenon has been primarily observed in Gd-based ferrimagnetic materials, which are unsuitable for storage due to their weak perpendicular magnetic anisotropy (PMA). In this work, we demonstrated that applying a single laser pulse and an in-plane magnetic field can facilitate magnetic switching in a Pt/Co/Pt ferromagnetic trilayers stack within a specific laser power window. To further understand this phenomenon, we introduce a Cu layer to accelerates the re-establishment time of the anisotropy field of Pt/Co/Pt trilayers, which leads to bullseye-patterned magnetic switching. We have mapped state diagrams for these phenomena, and through micromagnetic simulations, we have determined that these switchings are influenced by thermal anisotropy torque, which can be modulated through PMA. These findings indicate that single-shot optical precessional magnetization reversal is feasible in a broader range of materials, opening avenues for the development of optical-magnetic memory devices.

cond-mat.mtrl-sci↗

arXiv2025

Curriculum Learning for Biological Sequence Prediction: The Case of De Novo Peptide Sequencing

Peptide sequencing-the process of identifying amino acid sequences from mass spectrometry data-is a fundamental task in proteomics. Non-Autoregressive Transformers (NATs) have proven highly effective for this task, outperforming traditional methods. Unlike autoregressive models, which generate tokens sequentially, NATs predict all positions simultaneously, leveraging bidirectional context through unmasked self-attention. However, existing NAT approaches often rely on Connectionist Temporal Classification (CTC) loss, which presents significant optimization challenges due to CTC's complexity and increases the risk of training failures. To address these issues, we propose an improved non-autoregressive peptide sequencing model that incorporates a structured protein sequence curriculum learning strategy. This approach adjusts protein's learning difficulty based on the model's estimated protein generational capabilities through a sampling process, progressively learning peptide generation from simple to complex sequences. Additionally, we introduce a self-refining inference-time module that iteratively enhances predictions using learned NAT token embeddings, improving sequence accuracy at a fine-grained level. Our curriculum learning strategy reduces NAT training failures frequency by more than 90% based on sampled training over various data distributions. Evaluations on nine benchmark species demonstrate that our approach outperforms all previous methods across multiple metrics and species.

q-bio.BM↗

arXiv2025

Why Prompt Design Matters and Works: A Complexity Analysis of Prompt Search Space in LLMs

Despite the remarkable successes of large language models (LLMs), the underlying Transformer architecture has inherent limitations in handling complex reasoning tasks. Chain-of-thought (CoT) prompting has emerged as a practical workaround, but most CoT-based methods rely on a single, generic prompt such as "think step by step", with no task-specific adaptation. These approaches expect the model to discover an effective reasoning path on its own, forcing it to search through a vast prompt space. In contrast, several studies have explored task-specific prompt designs to boost performance. However, these designs are typically developed through trial and error, lacking theoretical grounding. As a result, prompt engineering remains largely ad hoc and unguided. In this paper, we provide a theoretical framework that explains why some prompts succeed while others fail. We show that prompts function as selectors, extracting task-relevant information from the model's full hidden state during CoT reasoning. Each prompt defines a unique trajectory through the answer space, and the choice of trajectory is crucial for task performance and future navigation within the space. We analyze the complexity of finding optimal prompts and characterize the size of the prompt space for a given task. Our theory reveals principles behind effective prompt design and shows that naive CoT-using self-guided prompts like "think step by step"-can severely hinder performance. Through experiments, we show that optimal prompt search can lead to more than a 50% improvement on reasoning tasks, providing a theoretical foundation for prompt engineering.

cs.CL↗

arXiv2025

Universal Biological Sequence Reranking for Improved De Novo Peptide Sequencing

De novo peptide sequencing is a critical task in proteomics. However, the performance of current deep learning-based methods is limited by the inherent complexity of mass spectrometry data and the heterogeneous distribution of noise signals, leading to data-specific biases. We present RankNovo, the first deep reranking framework that enhances de novo peptide sequencing by leveraging the complementary strengths of multiple sequencing models. RankNovo employs a list-wise reranking approach, modeling candidate peptides as multiple sequence alignments and utilizing axial attention to extract informative features across candidates. Additionally, we introduce two new metrics, PMD (Peptide Mass Deviation) and RMD (residual Mass Deviation), which offer delicate supervision by quantifying mass differences between peptides at both the sequence and residue levels. Extensive experiments demonstrate that RankNovo not only surpasses its base models used to generate training candidates for reranking pre-training, but also sets a new state-of-the-art benchmark. Moreover, RankNovo exhibits strong zero-shot generalization to unseen models whose generations were not exposed during training, highlighting its robustness and potential as a universal reranking framework for peptide sequencing. Our work presents a novel reranking strategy that fundamentally challenges existing single-model paradigms and advances the frontier of accurate de novo sequencing. Our source code is provided on GitHub.

cs.LG↗

arXiv2025

Tokenization Constraints in LLMs: A Study of Symbolic and Arithmetic Reasoning Limits

Tokenization is the first - and often underappreciated - layer of computation in language models. While Chain-of-Thought (CoT) prompting enables transformer models to approximate recurrent computation by externalizing intermediate steps, we show that the success of such reasoning is fundamentally bounded by the structure of tokenized inputs. This work presents a theoretical and empirical investigation into how tokenization schemes, particularly subword-based methods like byte-pair encoding (BPE), impede symbolic computation by merging or obscuring atomic reasoning units. We introduce the notion of Token Awareness to formalize how poor token granularity disrupts logical alignment and prevents models from generalizing symbolic procedures. Through systematic evaluation on arithmetic and symbolic tasks, we demonstrate that token structure dramatically affect reasoning performance, causing failure even with CoT, while atomically-aligned formats unlock strong generalization, allowing small models (e.g., GPT-4o-mini) to outperform larger systems (e.g., o1) in structured reasoning. Our findings reveal that symbolic reasoning ability in LLMs is not purely architectural, but deeply conditioned on token-level representations.

cs.CL↗

arXiv2021

Channel Modeling and Signal Processing for Array-based Visible Light Communication System in Misalignment

This paper proposes an indoor visible light communication (VLC) system with multiple transmitters and receivers. Due to diffusivity of LED light beams, photodiode receive signals from many directions. We use one concave and one convex lens as optical antenna, and obtain the optimal lens structure by optimizing which corresponds to the minimum condition number of channel gain matrix. In this way the light emitted by different LED can be separated well from each other then minimize signal interference. However, interference increases in the case of system deviation, so we explore the system mobility. Then subsequent signal processing is carried out, including signal combining and successive interference cancellation (SIC). We combine the same signal received by different receivers to improve signal to interference noise ratio (SINR). And SIC can effectively restore interference and eliminate its impact. The simulation results show that channel capacity can be increased by more than 5 times and up to 20 times under the condition of receiver and transmitter alignment. In the case of movement, channel capacity can also be increased by about 4 times on average. Moreover, the mobile range of system is also significantly expanded.

eess.SP↗

arXiv2020

All-optical helicity-independent switching state diagram in GdFeCo alloys

Ultra-fast magnetization switching induced by a single femtosecond laser pulse, under no applied magnetic field has attracted a lot of attention during the last 10 years because of its high potential for low energy and ultra-fast memory applications. Single-pulse helicity-independent switching has mostly been demonstrated for Gd based materials. It is now important to optimize the pulse duration and the energy needed to switch a GdFeCo magnet depending on the alloy thickness, concentration. Here we experimentally report state diagrams showing the magnetic state obtained after one single pulse depending on the laser pulse duration and fluence for various GdFeCo thin films with different compositions and thicknesses. We demonstrate that these state diagrams share similar characteristics: the fluence window for switching narrows for longer pulse duration and for the considered pulse duration range the critical fluence for single pulse switching increases linearly as a function of the pulse duration while the critical fluence required for creating a multidomain state remains almost constant. Calculations based on the atomistic spin model qualitatively reproduce the experimental state diagrams and their evolution. By studying the effect of the composition and the thickness on the state diagram, we demonstrated that the best energy efficiency and the longest pulse duration for switching are obtained for concentration around the magnetic compensation.

physics.optics↗

arXiv2020

Observation of Magnetic Droplets in Magnetic Tunnel Junctions

Magnetic droplets, a class of highly non-linear magnetodynamical solitons, can be nucleated and stabilized in nanocontact spin-torque nano-oscillators where they greatly increase the microwave output power. Here, we experimentally demonstrate magnetic droplets in magnetic tunnel junctions (MTJs). The droplet nucleation is accompanied by a power increase of over 300 times compared to its ferromagnetic resonance modes. The nucleation and stabilization of droplets are ascribed to the double-CoFeB free layer structure in the all-perpendicular MTJ which provides a low Zhang-Li torque and a high pinning field. Our results enable better electrical sensitivity in the fundamental studies of droplets and show that the droplets can be utilized in MTJ-based applications.

physics.app-ph↗

arXiv2018

Current-induced magnetization switching in atom-thick tungsten engineered perpendicular magnetic tunnel junctions with large tunnel magnetoresistance

Perpendicular magnetic tunnel junctions based on MgO/CoFeB structures are of particular interest for magnetic random-access memories because of their excellent thermal stability, scaling potential, and power dissipation. However, the major challenge of current-induced switching in the nanopillars with both a large tunnel magnetoresistance ratio and a low junction resistance is still to be met. Here, we report spin transfer torque switching in nano-scale perpendicular magnetic tunnel junctions with a magnetoresistance ratio up to 249% and a resistance area product as low as 7.0 Ω.μm2, which consists of atom-thick W layers and double MgO/CoFeB interfaces. The efficient resonant tunnelling transmission induced by the atom-thick W layers could contribute to the larger magnetoresistance ratio than conventional structures with Ta layers, in addition to the robustness of W layers against high temperature diffusion during annealing. The switching critical current density could be lower than 3.0 MA.cm-2 for devices with a 45 nm radius.

physics.app-ph↗