Search arXiv⌕ Search

arXiv subjects

Yu Zhu

Publications and source records attributed to Yu Zhu.

At least 19 recordsLinked to original sources
arXiv2026

Real-Time Prediction of Lower Limb Joint Kinematics, Kinetics, and Ground Reaction Force using Wearable Sensors and Machine Learning

Walking is a key movement of interest in biomechanics, yet gold-standard data collection methods are time- and cost-expensive. This paper presents a real-time, multimodal, high sample rate lower-limb motion capture framework, based on wireless wearable sensors and machine learning algorithms. Random Forests are used to estimate joint angles from IMU data, and ground reaction force (GRF) is predicted from instrumented insoles, while joint moments are predicted from angles and GRF using deep learning based on the ResNet-16 architecture. All three models achieve good accuracy compared to literature, and the predictions are logged at 1 kHz with a minimal delay of 23 ms for 20s worth of input data. The present work fully relies on wearable sensors, covers all five major lower limb joints, and provides multimodal comprehensive estimations of GRF, joint angles, and moments with minimal delay suitable for biofeedback applications.

eess.SY↗
arXiv2026

Environment-Aware Diffusion Model for Massive MIMO-OFDM Channel Estimation

This paper proposes an environment-aware diffusion based channel estimation in massive multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) systems. The high dimensionality of massive MIMO channels combined with limited pilot resources makes accurate estimation challenging. To address this issue, we exploit the spatial variability of wireless channels by training a diffusion model to learn the location-conditioned distribution of channel state information, which provides an environment-aware prior for channel estimation. Based on this learned prior, a posterior inference algorithm is developed to incorporate pilot observations into the reverse diffusion process, enabling Bayesian channel estimation by combining the received-signal likelihood with the learned channel prior. By jointly leveraging location information and measurement data, the proposed approach improves estimation accuracy under limited pilot resources. Simulation results based on ray-tracing channel datasets demonstrate that the proposed method consistently outperforms conventional estimators and existing learning-based approaches across various signal-to-noise ratios and pilot configurations.

eess.SP↗
arXiv2026

On Diagonalizable Delay-Doppler Channels and Their Diagonalizing Waveforms

In doubly selective channels, the joint delay and Doppler dispersion generally induces coupling among transmitted symbols, thereby increasing receiver equalization complexity. Nevertheless, by using appropriately designed waveforms, channels with certain delay-Doppler (DD) supports can be diagonalized for one-tap equalization. The whole picture of such DD supports and their corresponding waveforms is still unclear, except for several examples identified in literature. In this paper, under cyclic-prefix (CP)-based block transmission and assuming that the modulation waveforms form an orthonormal basis, we identify all such channel supports by an elementary expression, and derive the corresponding waveforms in closed form.

cs.IT↗
arXiv2026

Josephson energy of superconducting junctions: amorphous versus crystalline tunnel barriers

The Josephson energy $E_J$ is a key parameter governing the properties of transmon superconducting qubits. In Al/AlO$_x$/Al junctions, $E_J$ is set by electron tunneling through an ultrathin oxide barrier and therefore depends exponentially on the atomic structure of the barrier. We compute $E_J$ by first-principles device modeling based on the NEGF-DFT quantum-transport method, comparing a junction with a crystalline Al$_2$O$_3$ barrier against ten junctions with melt--quenched amorphous Al$_2$O$_3$ barriers of the same thickness. From the Fermi-level transmission and the Ambegaokar--Baratoff relation, we obtain a mean $E_J/h$ of $2.78$ GHz for the amorphous ensemble, with a standard deviation of $4.67$ GHz, compared with $0.73$ GHz for the crystalline reference; individual amorphous values span nearly two orders of magnitude. Scattering-state analysis shows that transport is quantum tunneling and that the variability originates from stoichiometric inhomogeneity of the amorphous oxide: Al-rich, low-barrier regions can connect into percolation-like tunneling pathways that strongly enhance the conductance. A realistic $200\times200$ nm$^2$ junction self-averages over more than $2\times10^4$ such microscopic regions. These results establish a quantitative atomistic route from oxide microstructure to the superconducting-circuit energy scale $E_J$.

quant-ph↗
arXiv2026

EasyBCI Agent: Towards Universal Neural Data Preprocessing for Brain-Computer Interfaces

Brain-computer interfaces translate neural activity into device commands, yet their performance hinges on preprocessing that remains manual, expert-dependent and poorly reproducible. Large language model agents can automate scientific coding, but existing systems lack the modality coverage, raw-data isolation, experience accumulation and domain oversight that neural preprocessing requires. We introduce EasyBCI, a two-phase LLM agent that plans and executes preprocessing pipelines for six signal types. A Plan Agent profiles each recording into a text-only Data Fingerprint that never exposes raw data to the model and selects a literature-grounded operator sequence. An Execution Agent generates, runs and self-corrects code until quality criteria are met, while a quality-gated experience system retains validated strategies as reusable skills. A domain expert intervenes at two decision gates, retaining human judgement where undetected error can invalidate downstream analyses. Evaluation on EEG with a fixed linear classifier shows that all five EasyBCI backbones preserve more task-relevant separability than the manual pipeline. Under same-backbone comparison, EasyBCI outperforms general-purpose coding agents on both label schemes for four of five configurations. EasyBCI extends to five additional modalities spanning nearly three orders of magnitude in sampling rate, producing complete reproducible pipelines with recorded decision provenance. These results indicate that domain-specific orchestration can bring auditable preprocessing within reach of laboratories lacking dedicated expertise, illustrating design principles applicable to AI agents in other scientific domains.

q-bio.QM↗
arXiv2026

Ontology-Guided Multi-Agent Extraction of Evaluation Objects from Academic Review Texts: Evidence from Chinese Library and Information Science

Academic reviews, scholarly commentaries, and book reviews serve as sources of evaluative statements about theories, methods, literature, institutions, and policies, providing valuable evidence for scholarly evaluation. Existing scientific entity extraction methods mainly target research articles and are less effective for evaluation objects, which are often abstract, context-dependent, and characterized by ambiguous type boundaries. This study proposes an ontology-guided multi-agent framework for evaluation object extraction. The framework combines candidate discovery, ontology-constrained classification, and domain review. Experimental results show that it achieves a Precision of 90.33%, Recall of 84.55%, Entity-level F1 of 87.34%, Strict Typed F1 of 79.78%, and Type Accuracy of 91.35%, substantially outperforming rule-based and zero-shot baselines. Ablation results indicate that the multi-agent workflow improves recall and stability, while ontology-based boundary constraints enhance fine-grained classification and reduce category confusion. The framework supports the structured utilization of evaluative scholarly texts and provides methodological support for evidence-based research evaluation and STI mining.

cs.CL↗
arXiv2026

Large-Small Model Collaboration for Zero-Shot Surgical Phase Recognition

Task-specific lightweight models for surgical phase recognition excel at capturing temporal dynamics but generalize poorly under domain shift. Conversely, surgical foundation models (FMs) offer superior transferability via large-scale pretraining, yet their lack of explicit temporal modeling often yields temporally inconsistent predictions, leading to degraded performance. To exploit the complementary strengths of both paradigms, we propose \textbf{La}rge-\textbf{S}mall \textbf{T}emporal adaptation (\textbf{LaST}), a novel large-small collaborative framework that enables zero-shot adaptation to unseen clinical domains. In LaST, the FM initiates the pipeline by generating frame-level phase priors that serve as initial weak supervision. To effectively utilize these noisy phase priors, we introduce an iterative temporal refinement scheme that integrates dynamic quality control to filter reliable predictions and dual-model cross-learning to mitigate confirmation bias. Simultaneously, the lightweight model leverages its intrinsic temporal modeling ability to progressively correct inconsistent predictions and enhance overall accuracy across iterations. At the end, a cycle replay strategy is employed to close the loop: the refined, more accurate predictions are utilized as upgraded supervision signals for the subsequent iterations, fostering a self-reinforcing evolution of both label quality and model capability. Extensive experiments demonstrate that LaST achieves robust adaptation to unseen domains for zero-shot surgical phase recognition, outperforming the baseline (PeskaVLP) by 24.85\%-43.17\% in accuracy and even surpassing fully supervised linear probing and several state-of-the-art few-shot approaches. Codes will be released at https://github.com/YIYIZH/LaST.

cs.CV↗
arXiv2026

Bayesian Federated Cause-of-Death Classification and Quantification Under Distribution Shift

In regions lacking medically certified causes of death, verbal autopsy (VA) is a widely used tool to ascertain the cause of death through interviews with caregivers. Data collected by VAs are often analyzed using probabilistic algorithms. The performance of these algorithms often degrades due to distribution shift across populations. Most existing VA algorithms rely on centralized training, requiring full access to training data for joint modeling. This can be infeasible due to privacy and logistical constraints. In this paper, we propose a novel Bayesian Federated Learning (BFL) framework that avoids data sharing across multiple training sources. Our method supports individual-level cause-of-death classification and population-level quantification of cause-specific mortality fractions in a target domain with limited or no local labeled data. The proposed framework is modular, computationally efficient, and compatible with a wide range of existing VA algorithms as base models, facilitating flexible deployment in real-world mortality surveillance systems. We validate the performance of BFL through extensive experiments on two real-world VA datasets under varying levels of distribution shift scenarios. Our results show that BFL significantly outperforms single-domain base models and performs comparably to or better than joint modeling.

stat.ME↗
arXiv2026

Extremal cross $t$-intersecting families under $t$-covering number constraints for vector spaces

Let $V$ be an $n$-dimensional vector space over the finite field $\mathbb{F}_q$, and ${V\brack k}$ denote the family of all $k$-dimensional subspaces of $V$. The families $\mathcal{F}\subseteq {V\brack k}$ and $\mathcal{G}\subseteq {V\brack \ell}$ are said to be cross $t$-intersecting if $\dim(F\cap G)\geq t$ for all $F\in\mathcal{F}$ and $G\in \mathcal{G}$. In this paper, we determine the extremal structures when $|\mathcal{F}||\mathcal{G}|$ attains the maximum value under the conditions $\dim\left(\cap_{F\in \mathcal{F}}F\right)<t$ and $\dim\left(\cap_{G\in \mathcal{G}}G\right)<t$.

math.CO↗
arXiv2026

SeekBrain: An Autonomous Multi-Agent System for Accelerating Neuroscience Discovery

Modern neuroscience relies on integrating multi-scale, multimodal datasets to uncover the neural principles underlying intelligence. However, analytical challenges posed by highly heterogeneous data and fragmented workflows increasingly constrain discoveries. Here we introduce SeekBrain, an autonomous multi-agent framework designed to accelerate neuroscience discovery through domain-grounded hierarchical planning and cross-modal data analysis. SeekBrain dynamically constructs a repertoire of analysis recipes extracted from code-paper pairs. By coupling this codified expertise with agentic planning and execution engines, the framework scalably generates hypotheses and analytical pipelines on demand. Systematic evaluation on the expert-annotated BrainArena benchmark demonstrates that SeekBrain substantially outperforms state-of-the-art agent baselines across various analysis tasks. Crucially, when deployed in real-world research, SeekBrain integrated behavioral, neural, and anatomical data to reveal structured, distributed neural representations of larval zebrafish behavior and a shared axis of regional decoding strength across the brain in a mouse decision-making task. These results establish SeekBrain as a scalable and practical tool for accelerating data-driven discoveries in neuroscience.

cs.MA↗
arXiv2026

Application-Driven Architecture Exploration for Cross-Layer Heterogeneous Systems

AI and HPC infrastructure increasingly serves workload portfolios that combine dense tensor computation, sparse kernels, large memory footprints, and communication-intensive collectives. Supporting these portfolios requires coordinated choices across accelerators, memory tiers, scale-up fabrics, and cluster networks. The resulting Cross-layer Heterogeneous System (XHS) design space is difficult to explore: hardware choices change legal task mappings, while rack power, switch radix, cabling, and cost constraints invalidate many candidates. We present CHASE, an application-driven framework that searches physically feasible XHS architectures through the workloads they must execute. CHASE represents candidates as hierarchical typed graphs and rejects designs that violate deployment constraints. It avoids intractable joint hardware-mapping search with a decoupled two-level loop: an inner mapper translates hardware-independent workload DAGs into topology-aware event traces, a calibrated event-driven simulator evaluates each mapping, and an outer telemetry-guided optimizer evolves the hardware graph. We evaluate CHASE on sparse-computing and LLM workloads. Its mapper remains within 6.06% of exhaustive optima while reducing mapping time by 60.5% on average relative to PEFT. Compute-model errors average 4.4-7.5%, and communication validation reproduces key trends across physical platforms. The outer search reaches near-global optima within 64 iterations. End-to-end case studies show that sparse workloads favor criticality-aware heterogeneous pods, whereas LLM inference favors scale-up islands; the resulting designs deliver 6.20$\times$ and 2.12$\times$ geomean speedups, respectively, while reducing cost and power relative to the baselines.

cs.DC↗
arXiv2026

First-principles predictions of band alignment in strained Si/Si1-xGex and Ge/Si1-xGex heterostructures

Accurate band offsets are essential for predictive continuum modeling of nanostructures such as quantum wells and quantum dots formed in strained Si/Si1-xGex and Ge/Si1-xGex heterostructures. Experimental offset data for these systems remain sparse away from endpoint compositions, making composition-dependent design difficult. We use atomistic first-principles density functional theory to compute valence- and conduction-band offsets across the full range 0 <= x <= 1. Random alloying is treated with special quasirandom structures, interface lineup terms are extracted from macroscopically averaged local Kohn-Sham potentials in thick periodic superlattices, valence-band spin-orbit coupling is included through species-resolved Mulliken weights, and conduction-band edges are refined using the screened hybrid Heyd-Scuseria-Ernzerhof functional. The resulting offsets show pronounced composition nonlinearity beyond the linear models explored in previous works, agree with experimental benchmarks, and reproduce the high-Ge slope change in the relaxed-alloy band gap. Analytic fitting expressions are provided for direct use in simulations, facilitating practical design of modern quantum technology devices.

cond-mat.mes-hall↗
arXiv2026

Food-R1: A Unified Multi-Task Food Vision-Language Model with Reinforcement Learning

Recent studies have explored Vision-Language Models (VLMs) for food analysis. However, most existing methods rely primarily on supervised fine-tuning (SFT), which often limits reasoning and generalization capabilities. Moreover, high-quality large-scale nutritional annotations remain scarce. To address these issues, we introduce CalorieBench-80K, a large-scale benchmark with curated calorie labels and dietary advice annotations. To the best of our knowledge, it is the first food image benchmark to incorporate Chain-of-Thought (CoT) annotations for calorie reasoning. We also propose Food-R1, a unified food VLM trained in a multi-task learning paradigm to equip the model with broad capabilities. Food-R1 undergoes CoT-based cold-start instruction tuning, followed by reinforcement fine-tuning (RFT) using Group Relative Policy Optimization (GRPO) to improve reasoning and performance. Experiments on CalorieBench-80K and representative benchmarks show that Food-R1 consistently outperforms strong baselines across food-related tasks. The code, model weights, and benchmark annotations are available at the project repository.

cs.CV↗
arXiv2026

MatterSim-MT: A multi-task foundation model for in silico materials characterization

Accurate property characterization is a major bottleneck in materials design. While first-principles methods and task-specific machine-learning models have driven important progress, they remain fundamentally limited in scalability and generalizability across the vast space of structures and properties relevant to real-world materials design. We present MatterSim-MT, a multi-task foundation model for in silico materials simulation and property characterization. The model is pretrained on over 35 million first-principles-labeled structures covering 89 elements, temperatures up to 5000 K and pressures up to 1000 GPa, and is fine-tuned on various properties including Bader charges, magnetic moments, Born effective charges, and dielectric matrices. Out of the box, MatterSim-MT not only serves as a foundation model for predicting material structure, dynamics and thermodynamics, its multi-task architecture also enables a wide range of complex simulations that cannot be captured by potential energy surfaces alone. For example, we demonstrate pressure-dependent LO-TO phonon splitting in SiC with close agreement with experiment, electric hysteresis in ferroelectric BaTiO3, and the cationic-to-anionic redox transition during delithiation of a Li-rich cathode material. Finally, we show that MatterSim-MT scales well with more data and parameters, can be efficiently fine-tuned to higher levels of theory, and can be efficiently extended to new systems via active learning. Overall, we believe this approach provides a scalable route to accurate in silico materials characterization.

cond-mat.mtrl-sci↗
arXiv2026

Equip Pre-ranking with Target Attention by Residual Quantization

The pre-ranking stage in industrial recommendation systems faces a fundamental conflict between efficiency and effectiveness. While powerful models like Target Attention (TA) excel at capturing complex feature interactions in the ranking stage, their high computational cost makes them infeasible for pre-ranking, which often relies on simplistic vector-product models. This disparity creates a significant performance bottleneck for the entire system. To bridge this gap, we propose TARQ, a novel pre-ranking framework. Inspired by generative models, TARQ's key innovation is to equip pre-ranking with an architecture approximate to TA by Residual Quantization. This allows us to bring the modeling power of TA into the latency-critical pre-ranking stage for the first time, establishing a new state-of-the-art trade-off between accuracy and efficiency. Extensive offline experiments and large-scale online A/B tests at Taobao demonstrate TARQ's significant improvements in ranking performance. Consequently, our model has been fully deployed in production, serving tens of millions of daily active users and yielding substantial business improvements. The code and data are available at https://github.com/zyody/tarq_sigir2026.

cs.IR↗
arXiv2026

Analysis of Hydrogen Contamination in Al/AlOx/Al Josephson Junctions

Hydrogen contamination in Josephson junctions is a potential source of device-to-device variability and two-level-system loss in superconducting qubits. In this work, we investigate hydrogen incorporation in oxidized aluminum barriers by combining molecular dynamics simulations with atomistic quantum transport calculations. The oxide growth simulations are performed using CHGNet for Al surfaces exposed to dense O$_{\text{2}}$ and H$_{\text{2}% }$O environments, yielding amorphous AlO$_{\text{x}}$ layers with hydrogen content comparable to experimentally relevant levels. From $400$ statistically independent samples, we find that the number of H atoms in the oxide is well described by a beta-binomial distribution, reflecting correlations induced by the self-limiting oxidation process. Structural analysis shows that most hydrogen atoms reside near the AlO$_{\text{x}}$ surface and predominantly form Al-OH and Al-OH-Al motifs. To assess the impact of hydrogen on transport, we construct Al/Al$_{\text{2}}$O$_{\text{3}} $/Al junction models and perform NEGF-DFT calculations with NanoDCAL, using a GGA+U scheme to calibrate the band gap and band alignment. H atoms are found to increase the transmission coefficient near the Fermi level and shift the electronic structure in a manner consistent with effective p-type doping. By combining the H atom number statistics from molecular dynamics with the transmission coefficients from quantum transport calculations, we obtain a probability distribution for the Josephson energy. For a Josephson junction with an average hydrogen content of $2.56$ at.\%, the resulting Josephson energy is predicted to be $% E_{J}/h=10.92\pm 0.26$ GHz. These results provide an atomistic picture of hydrogen contamination and an estimate of device variability in Josephson junctions.

quant-ph↗
arXiv2026

On extremal cross $t$-intersecting families with $t$-covering number conditions

Let $n$, $k$ and $t$ be positive integers, and let $\mathcal{F}$ be a collection of $k$-subsets of $[n]=\{1,2,\dots,n\}$. The $t$-covering number $τ_t(\mathcal{F})$ of $\mathcal{F}$ is defined as the minimum size of a set $T$ such that $|F\cap T|\geq t$ for all $F\in \mathcal{F}$. For positive integers $k_1$ and $k_2$, let $\mathcal{F}_i$ be a collection of $k_i$-subsets of $[n]$ for $i\in \{1,2\}$. The families $\mathcal{F}_1$ and $\mathcal{F}_2$ are said to be cross $t$-intersecting if $|F_1\cap F_2|\geq t$ for all $F_1\in\mathcal{F}_1$ and $F_2\in \mathcal{F}_2$. When $\mathcal{F}_1=\mathcal{F}_2$, $\mathcal{F}_1$ is called a $t$-intersecting family. In this paper, we first characterize the extremal structures of cross $t$-intersecting families $\mathcal{F}_1$ and $\mathcal{F}_2$ that maximize $|\mathcal{F}_1||\mathcal{F}_2|$ under the condition that $τ_t(\mathcal{F}_1)\geq t+1$ and $τ_t(\mathcal{F}_2)\geq t+1$. We then describe the maximal $t$-intersecting families with $t$-covering number $t+1$.

math.CO↗
arXiv2026

Early Stopping Chain-of-thoughts in Large Language Models

Reasoning large language models (LLMs) have demonstrated superior capacities in solving complicated problems by generating long chain-of-thoughts (CoT), but such a lengthy CoT incurs high inference costs. Previous methods on inference-stage efficient reasoning either require white-box models to monitor the reasoning process or are not reliable through direct prompting. In response, we introduce ES-CoT, an inference-time method that shortens CoT generation by detecting answer convergence and stopping early with almost no performance loss. When observing a linguistic marker (such as "wait") in the reasoning process, we prompt the LLM to output its current final answer, denoted as a step answer. We then track the run length of consecutive identical step answers as a measure of answer convergence. We show both empirically and theoretically that step answers steadily converge to the final answer, and large run-length jumps reliably mark this convergence. Experiments on six reasoning datasets across three LLMs show that ES-CoT reduces the number of inference tokens by 16.08% on average while maintaining accuracy comparable to standard CoT.

cs.CL↗