Researcher profile

Yuxuan Chen

Yuxuan Chen contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate

Computer Vision Computation and Language Artificial Intelligence cond-mat.mtrl-sci eess.IV Information Theory Machine Learning math.IT Multimedia Robotics

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author

9works

0followers

10topics

4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Dynamic Execution Commitment of Vision-Language-Action Models

Vision-Language-Action (VLA) models predominantly adopt action chunking, i.e., predicting and committing to a short horizon of consecutive low-level actions in a single forward pass, to amortize the inference cost of large-scale backbones and reduce per-step latency. However, committing these multi-step predictions to real-world execution requires balancing success rate against inference efficiency, a decision typically governed by fixed execution horizons tuned per task. Such heuristics ignore the state-dependent nature of predictive reliability, leading to brittle performance in dynamic or out-of-distribution settings. In this paper, we introduce A3, an Adaptive Action Acceptance mechanism that reframes dynamic execution commitment as a self-speculative prefix verification problem. A3 first computes a trajectory-wise consensus score of actions via group sampling, then selects a representative draft and prioritizes downstream verification. Specifically, it enforces: (1) consensus-ordered conditional invariance, which validates low-consensus actions by judging whether they remain consistent when re-decoded conditioned on high-consensus actions; and (2) prefix-closed sequential consistency, which guarantees physical rollout integrity by accepting only the longest continuous sequence of verified actions starting from the beginning. Consequently, the execution horizon emerges as the longest verifiable prefix satisfying both internal model logic and sequential execution constraints. Experiments across diverse VLA models and benchmarks demonstrate that A3 eliminates the need for manual horizon tuning while achieving a superior trade-off between execution robustness and inference throughput.

preprint2026arXiv

Engineering Favorable Propagation: Near-Field IRS Deployment for Spatial Multiplexing

In intelligent reflecting surface IRS assisted multiple input multiple output MIMO systems, a strong line of sight LoS link is required to compensate for the severe cascaded path loss. However, such a link renders the effective channel highly rank deficient and fundamentally limits spatial multiplexing. To overcome this limitation, this paper leverages the large aperture of sparse arrays to harness near field spherical wavefronts, and establishes a deterministic deployment criterion that strategically positions the IRS in the near field of a base station BS. This placement exploits the spherical wavefronts of the BS IRS link to engineer decorrelated channels, thereby fundamentally overcoming the rank deficiency issue in far field cascaded channels. Based on a physical channel model for the sparse BS array and the IRS, we characterize the rank properties and inter user correlation of the cascaded BS IRS user channel. We further derive a closed form favorable propagation metric that reveals how the sparse array geometry and the IRS position can be tuned to reduce inter user channel correlation. The resulting geometry driven deployment rule provides a simple guideline for creating a favorable propagation environment with enhanced effective degrees of freedom. The favorable channel statistics induced by our deployment criterion enable a low complexity maximum ratio transmission MRT precoding scheme. This serves as the foundation for an efficient algorithm that jointly optimizes the IRS phase shifts and power allocation based solely on long term statistical channel state information CSI. Simulation results validate the effectiveness of our deployment criterion and demonstrate that our optimization framework achieves significant performance gains over benchmark schemes.

preprint2026arXiv

External Validation of Deep Learning Models for BI-RADS Breast Density Prediction from Ultrasound Images

We externally validated three deep learning models (DenseNet121, ViT-B/32, and ResNet50) for predicting mammographic breast density from breast ultrasound exams on an independent cohort. The external validation set comprised 2,000 ultrasound exams, including 500 cancer cases defined by an initial negative exam (BI-RADS 1 or 2) followed by a cancer diagnosis within 6 months to 10 years, and 1,500 negative controls matched by manufacturer and study year. Performance was measured using patient-level AUROC across four density categories: A (fatty), B (scattered), C (heterogeneous), and D (extremely dense). As a downstream assessment, we also evaluated 10-year risk prediction by incorporating age and AI-derived density into the Tyrer-Cuzick model and comparing performance against a reference model using age and mammography-reported density. All three models performed best in extremely dense breasts (AUROC 0.868-0.899), with strong performance in fatty (0.814-0.838) and scattered density (0.764-0.799), and lower performance in heterogeneously dense breasts (0.699-0.729). DenseNet121 achieved the highest overall performance (micro-averaged AUROC 0.885), and performance across categories was comparable between internal and external testing. For risk modeling, age combined with AI-derived density yielded a lower AUROC than age combined with mammography-reported density (0.541 vs. 0.570; p = 0.23), with no statistically significant difference. These findings indicate that deep learning models generalize well to external data with different racial composition for breast density assessment. While performance is strongest in extremely dense breasts, heterogeneously dense remains more challenging, highlighting the need for targeted optimization.

preprint2026arXiv

PROVE: A Perceptual RemOVal cohErence Benchmark for Visual Media

Evaluating object removal in images and videos remains challenging because the task is inherently one-to-many, yet existing metrics frequently disagree with human perception. Full-reference metrics reward copy-paste behaviors over genuine erasure; no-reference metrics suffer from systematic biases such as favoring blurry results; and global temporal metrics are insensitive to localized artifacts within edited regions. To address these limitations, we propose RC (Removal Coherence), a pair of perception-aligned metrics: RC-S, which measures spatial coherence via sliding-window feature comparison between masked and background regions, and RC-T, which measures temporal consistency via distribution tracking within shared restored regions across adjacent frames. To validate RC and support community benchmarking, we further introduce PROVE-Bench, a two-tier real-world benchmark comprising PROVE-M, an 80-video paired dataset with motion augmentation, and PROVE-H, a 100-video challenging subset without ground truth. Together, RC metrics and PROVE-Bench form the PROVE (Perceptual RemOVal cohErence) evaluation framework for visual media. Experiments across diverse image and video benchmarks demonstrate that RC achieves substantially stronger alignment with human judgments than existing evaluation protocols. The code for RC metrics and PROVE-Bench are publicly available at: https://github.com/xiaomi-research/prove/.

preprint2022arXiv

A Comparative Study of Pre-trained Encoders for Low-Resource Named Entity Recognition

Pre-trained language models (PLM) are effective components of few-shot named entity recognition (NER) approaches when augmented with continued pre-training on task-specific out-of-domain data or fine-tuning on in-domain data. However, their performance in low-resource scenarios, where such data is not available, remains an open question. We introduce an encoder evaluation framework, and use it to systematically compare the performance of state-of-the-art pre-trained representations on the task of low-resource NER. We analyze a wide range of encoders pre-trained with different strategies, model architectures, intermediate-task fine-tuning, and contrastive learning. Our experimental results across ten benchmark NER datasets in English and German show that encoder performance varies significantly, suggesting that the choice of encoder for a specific low-resource scenario needs to be carefully evaluated.

preprint2022arXiv

Creating a Nanoscale Lateral Heterojunction in a Semiconductor Monolayer with a Large Built-in Potential

The ability to engineer atomically thin nanoscale lateral heterojunctions (HJs) is critical to lay the foundation for future two-dimensional (2D) device technology. However, the traditional approach to creating a heterojunction by direct growth of a heterostructure of two different materials constrains the available band offsets, and it is still unclear if large built-in potentials are attainable for 2D materials. The electronic properties of atomically thin semiconducting transition metal dichalcogenides (TMDs) are not static, and their exciton binding energy and quasiparticle band gap depend strongly on the proximal environment. Recent studies have shown that this effect can be harnessed to engineer the lateral band profile of monolayer TMDs to create a heterojunction. Here we demonstrate the synthesis of a nanoscale lateral heterojunction in monolayer MoSe2 by intercalating Se at the interface of a hBN/Ru(0001) substrate. The Se intercalation creates a spatially abrupt modulation of the local hBN/Ru work function, which is imprinted directly onto an overlying MoSe2 monolayer to create a large built-in potential of 0.83 eV. We spatially resolve the MoSe2 band profile and work function using scanning tunneling spectroscopy to map out the nanoscale depletion region. The Se intercalation also modifies the dielectric environment, influencing the local band gap renormalization and increasing the MoSe2 band gap by ~0.26 eV. This work illustrates that environmental proximity engineering provides a robust method to indirectly manipulate the band profile of 2D materials outside the limits of their intrinsic properties, providing avenues for future device design.

preprint2022arXiv

SLIC: Self-Supervised Learning with Iterative Clustering for Human Action Videos

Self-supervised methods have significantly closed the gap with end-to-end supervised learning for image classification. In the case of human action videos, however, where both appearance and motion are significant factors of variation, this gap remains significant. One of the key reasons for this is that sampling pairs of similar video clips, a required step for many self-supervised contrastive learning methods, is currently done conservatively to avoid false positives. A typical assumption is that similar clips only occur temporally close within a single video, leading to insufficient examples of motion similarity. To mitigate this, we propose SLIC, a clustering-based self-supervised contrastive learning method for human action videos. Our key contribution is that we improve upon the traditional intra-video positive sampling by using iterative clustering to group similar video instances. This enables our method to leverage pseudo-labels from the cluster assignments to sample harder positives and negatives. SLIC outperforms state-of-the-art video retrieval baselines by +15.4% on top-1 recall on UCF101 and by +5.7% when directly transferred to HMDB51. With end-to-end finetuning for action classification, SLIC achieves 83.2% top-1 accuracy (+0.8%) on UCF101 and 54.5% on HMDB51 (+1.6%). SLIC is also competitive with the state-of-the-art in action classification after self-supervised pretraining on Kinetics400.

preprint2022arXiv

Why only Micro-F1? Class Weighting of Measures for Relation Classification

Relation classification models are conventionally evaluated using only a single measure, e.g., micro-F1, macro-F1 or AUC. In this work, we analyze weighting schemes, such as micro and macro, for imbalanced datasets. We introduce a framework for weighting schemes, where existing schemes are extremes, and two new intermediate schemes. We show that reporting results of different weighting schemes better highlights strengths and weaknesses of a model.

preprint2021arXiv

PiP: Planning-informed Trajectory Prediction for Autonomous Driving

It is critical to predict the motion of surrounding vehicles for self-driving planning, especially in a socially compliant and flexible way. However, future prediction is challenging due to the interaction and uncertainty in driving behaviors. We propose planning-informed trajectory prediction (PiP) to tackle the prediction problem in the multi-agent setting. Our approach is differentiated from the traditional manner of prediction, which is only based on historical information and decoupled with planning. By informing the prediction process with the planning of ego vehicle, our method achieves the state-of-the-art performance of multi-agent forecasting on highway datasets. Moreover, our approach enables a novel pipeline which couples the prediction and planning, by conditioning PiP on multiple candidate trajectories of the ego vehicle, which is highly beneficial for autonomous driving in interactive scenarios.

Institution

Affiliation not imported yet

This author record came from a source that does not expose affiliation metadata. Once the author claims the profile or we enrich the record from another provider, this section will link to the concrete institution.

Topic footprint