arXiv subjects

Jialiang Chen

Publications and source records attributed to Jialiang Chen.

12 recordsLinked to original sources

arXiv2026

AOI-Net: Structural Face AOI-Guided Eye-Gaze Track Representation Learning for Autism Spectrum Disorder Detection

Eye-movement tracking has emerged as a promising non-invasive approach to Autism Spectrum Disorder (ASD) screening, with systematic differences in attentional allocation and revisit behaviors observed during socially interactive tasks. Existing computational methods typically characterize eye-movements using discrete gaze trajectories and fixation events, yielding representations dominated by short-range temporal dynamics and limiting models that primarily emphasize long-range dependencies. Meanwhile, gaze behavior is naturally organized across semantically meaningful Areas of Interest (AOIs), whose attention allocation and transitions provide important structural cues, yet their relationships are rarely modeled explicitly. To address these limitations, we propose a structural face AOI-guided Eye-Gaze Track Network (AOI-Net) that jointly models short-term temporal dynamics and AOI-level structural organization. A network gating mechanism adaptively integrates the complementary temporal and structural representations according to their contributions to gaze-behavior characterization. To mitigate the pronounced class imbalance commonly encountered between individuals with ASD and Typically Developing (TD) participants in clinical datasets, class-distribution-aware learning is further employed to facilitate discriminative embedding learning under skewed class distributions. Experiments on a unique and large-scale clinical eye-tracking database comprising eight stimulus subsets and more than 1,300 participants show that AOI-Net consistently outperforms state-of-the-art methods. The proposed framework also enables interpretable gaze-behavior modeling and provides a practical basis for scalable AI-driven ASD screening in real-world healthcare. The code is available at https://github.com/Zhanpei-ai/CIM-AOI-Net/tree/main/Code

cs.CV↗

arXiv2026

Robust Activation Map Rectification for Weakly Supervised Volumetric Segmentation: Temporal Coherence as a Free Lunch

Weakly supervised segmentation relies heavily on class activation maps (CAMs) to initially localize target regions. However, CAMs are often noisy and prone to catastrophic failures. Existing remedies typically introduce additional training stages or prototype learning, increasing computational cost and reducing robustness. In this paper, we propose a training-free prototype-free framework that rectifies unreliable CAMs by exploiting temporal and structural coherence in volumetric data as a free lunch. Our approach is built on two key components. First, we introduce Variance-Reduced Activation Aggregation (VRAA) which suppresses noise and amplify coherent semantic signals. We provide a theoretical justification by modeling CAMs as high-dimensional random vectors and show that aggregation yields provable variance reduction. Second, we design a Bidirectional Extremity Rectification (BER) mechanism that detects and rectifies implausible activations through bidirectional extremity checks, effectively mitigating extreme-value failures without learning additional parameters. Our method is model-agnostic and can be seamlessly integrated with existing pipelines. Extensive experiments on multiple public benchmarks demonstrate substantial improvements over state-of-the-art weakly supervised methods, achieving up to 20% Dice and 40% mIoU gains while reducing inference time by more than 5 times. These results indicate that leveraging coherence as an implicit inductive bias yields a principled and efficient approach to stabilizing weakly supervised volumetric segmentation. Our code will be available.

cs.CV↗

arXiv2026

World Narrative Model for Highly Controllable Video Generation: A Paradigm Shift from Pixel Sampling to Physical World Orchestration

The fundamental obstacle to industrial grade video generation is the lack of controllability: existing models treat video as a pixel distribution sampling problem, bypassing the explicit, instance level $4D$ $(3D + T)$ physical world. Consequently, content creators cannot specify geometry, motion, camera parameters, or lighting in a deterministic, quantitative way, leading to the infamous ''gacha'' loop that makes professional content creation prohibitively inefficient and expensive. To address this, we introduce the World Narrative Model (WNM), a paradigm that decouples what to render -- the structured physical narrative -- from how to render -- the pixel generation process. WNM replaces end-to-end black-box sampling with orchestrated $4D$ pre-visualization for media generation. Collaborative agents translate sparse multimodal inputs, including text, reference videos, and sketches, into a fully editable world representation with scene geometry, object layouts, character/animal skeleton motion, trajectories, camera motion, and lighting at quantitative, physically meaningful granularity. This representation acts as a deterministic structural blueprint that drives existing video foundation models, either frozen or lightly adapted, to render final footage, turning the base model into a faithful neural shader. Built on this engine, our human-AI platform supports automatic world generation and pre-visualization aligned with professional filmmaking pipelines, while director consoles enable seamless human refinement. Experiments show that WNM greatly reduces probabilistic ``gacha'' calls and produces videos whose layout, motion, and cinematography closely follow creator intent. The framework is open and modular, allowing each component, such as world representation, control agents, and adapters, to be independently improved. Project website: https://glassroom.sjtu.edu.cn/WNM/.

cs.CV↗

arXiv2026

DIRECT: Video Mashup Creation via Hierarchical Multi-Agent Planning and Intent-Guided Editing

Video mashup creation represents a complex video editing paradigm that recomposes existing footage to craft engaging audio-visual experiences, demanding intricate orchestration across semantic, visual, and auditory dimensions and multiple levels. However, existing automated editing frameworks often overlook the cross-level multimodal orchestration to achieve professional-grade fluidity, resulting in disjointed sequences with abrupt visual transitions and musical misalignment. To address this, we formulate video mashup creation as a Multimodal Coherency Satisfaction Problem (MMCSP) and propose the DIRECT framework. Simulating a professional production pipeline, our hierarchical multi-agent framework decomposes the challenge into three cascade levels: the Screenwriter for source-aware global structural anchoring, the Director for instantiating adaptive editing intent and guidance, and the Editor for intent-guided shot sequence editing with fine-grained optimization. We further introduce Mashup-Bench, a comprehensive benchmark with tailored metrics for visual continuity and auditory alignment. Extensive experiments demonstrate that DIRECT significantly outperforms state-of-the-art baselines in both objective metrics and human subjective evaluation. Project page and code: https://github.com/AK-DREAM/DIRECT

cs.CV↗

arXiv2025

Sulfur monoxide as a tracer of the Galactic $^{32}$S/$^{34}$S gradient

To date, the Galactic interstellar radial $^{32}$S/$^{34}$S gradient has only been studied with the CS isotopologs, which may be affected by uncertainties due to the use of a single tracer. As another abundant S-bearing molecules, SO and its isotopomer $^{34}$SO could be considered as tracers of the $^{32}$S/$^{34}$S ratio. We present the first systematic observations of SO and $^{34}$SO toward a large sample of molecular clouds with accurate distances, performed with the IRAM 30 m and the 10 m Submillimeter Telescope (SMT). With the IRAM 30 m, SO $2_2-1_1$ was detected in 59 of 82 sources ($\sim$82%), and $^{34}$SO $2_2-1_1$ in 8 sources ($\sim$10%). With the SMT 10 m, SO $5_5-4_4$ was detected in 136 of 184 sources ($\sim$74%), and $^{34}$SO $5_5-4_4$ in 55 of 77 strong SO sources ($\sim$72%). SO/$^{34}$SO ratios were derived for 8 ($2_2-1_1$) and 55 ($5_5-4_4$) sources. No correlation was found between the SO/$^{34}$SO ratio and heliocentric distance or $T_k$, suggesting negligible distance and fractionation effects. Both LTE and non-LTE analyses consistently suggest that the optical depth effect is also insignificant. $^{32}$S/$^{34}$S ratios from the $2_2-1_1$ transitions follow the increasing radial trend proposed by previous CS species measurements, while those from the $5_5-4_4$ lines are systematically lower. The lower transitions of SO and $^{34}$SO may be suitable tracers of $^{32}$S/$^{34}$S, though the detections are rare. Comparisons between measurements and Galactic chemical evolution model suggest that the nucleosynthesis prescriptions need to be revised in the low-metallicity regime, but more data for the outermost Galactic regions are crucial for drawing strong conclusions.

astro-ph.GA↗

arXiv2025

CMD: Controllable Multiview Diffusion for 3D Editing and Progressive Generation

Recently, 3D generation methods have shown their powerful ability to automate 3D model creation. However, most 3D generation methods only rely on an input image or a text prompt to generate a 3D model, which lacks the control of each component of the generated 3D model. Any modifications of the input image lead to an entire regeneration of the 3D models. In this paper, we introduce a new method called CMD that generates a 3D model from an input image while enabling flexible local editing of each component of the 3D model. In CMD, we formulate the 3D generation as a conditional multiview diffusion model, which takes the existing or known parts as conditions and generates the edited or added components. This conditional multiview diffusion model not only allows the generation of 3D models part by part but also enables local editing of 3D models according to the local revision of the input image without changing other 3D parts. Extensive experiments are conducted to demonstrate that CMD decomposes a complex 3D generation task into multiple components, improving the generation quality. Meanwhile, CMD enables efficient and flexible local editing of a 3D model by just editing one rendered image.

cs.CV↗

arXiv2025

A Resource-Efficient Training Framework for Remote Sensing Text--Image Retrieval

Remote sensing text--image retrieval (RSTIR) aims to retrieve the matched remote sensing (RS) images from the database according to the descriptive text. Recently, the rapid development of large visual-language pre-training models provides new insights for RSTIR. Nevertheless, as the complexity of models grows in RSTIR, the previous studies suffer from suboptimal resource efficiency during transfer learning. To address this issue, we propose a computation and memory-efficient retrieval (CMER) framework for RSTIR. To reduce the training memory consumption, we propose the Focus-Adapter module, which adopts a side branch structure. Its focus layer suppresses the interference of background pixels for small targets. Simultaneously, to enhance data efficacy, we regard the RS scene category as the metadata and design a concise augmentation technique. The scene label augmentation leverages the prior knowledge from land cover categories and shrinks the search space. We propose the negative sample recycling strategy to make the negative sample pool decoupled from the mini-batch size. It improves the generalization performance without introducing additional encoders. We have conducted quantitative and qualitative experiments on public datasets and expanded the benchmark with some advanced approaches, which demonstrates the competitiveness of the proposed CMER. Compared with the recent advanced methods, the overall retrieval performance of CMER is 2%--5% higher on RSITMD. Moreover, our proposed method reduces memory consumption by 49% and has a 1.4x data throughput during training. The code of the CMER and the dataset will be released at https://github.com/ZhangWeihang99/CMER.

cs.CV↗

arXiv2025

CoReQA: Uncovering Potentials of Language Models in Code Repository Question Answering

Large language models that enhance software development tasks, such as code generation, code completion, and code question answering (QA), have been extensively studied in both academia and the industry. The models are integrated into popular intelligent IDEs like JetBrains and Cursor. Current benchmarks for evaluating models' code comprehension capabilities primarily focus on code generation or completion, often neglecting QA, which is a crucial aspect of understanding code. Existing code QA benchmarks are derived from code comments with predefined patterns (e.g., CodeQA) or focus on specific domains, such as education (e.g., CS1QA). These benchmarks fail to capture the real-world complexity of software engineering and user requirements for understanding code repositories. To address this gap, we introduce CoReQA, a benchmark for Code Repository-level question answering, constructed from GitHub issues and comments from 176 popular repositories across four programming languages. Since questions and answers may include both natural language and code snippets, traditional evaluation metrics such as BLEU are inadequate for assessing repository-level QA performance. Thus, we provide an LLM-as-a-judge framework to evaluate QA performance from five aspects. Based on CoReQA, we evaluate the performance of three baselines, including two short-context models using generic retrieval strategies and one long-context model that utilizes the entire repository context. Evaluation results show that state-of-the-art proprietary and long-context models struggle to address repository-level questions effectively. Our analysis highlights the limitations of language models in assisting developers in understanding repositories and suggests future directions for improving repository comprehension systems through effective context retrieval methodologies.

cs.SE↗

arXiv2024

Exploring Parameter-Efficient Fine-Tuning of Large Language Model on Automated Program Repair

Automated Program Repair (APR) aims to fix bugs by generating patches. And existing work has demonstrated that "pre-training and fine-tuning" paradigm enables Large Language Models (LLMs) improve fixing capabilities on APR. However, existing work mainly focuses on Full-Model Fine-Tuning (FMFT) for APR and limited research has been conducted on the execution-based evaluation of Parameter-Efficient Fine-Tuning (PEFT) for APR. Comparing to FMFT, PEFT can reduce computing resource consumption without compromising performance and has been widely adopted to other software engineering tasks. To fill this gap, we enhance the existing APR dataset by employing prompt engineering to create an instruction dataset, APR-INSTRUCTION, at first. Secondly, we fine-tune four pre-trained LLMs using four different PEFT methods with APR-INSTRUCTION. The best fine-tuned model fixes 58% more bugs than the state-of-the-art LLM-based APR techniques. The results also show that $(IA)^3$ improves the creativity of LLMs more effectively through fine-tuning and achieves the highest fixing capability compared to the other three PEFT methods. Thirdly, we explore the optimal configuration of PEFT hyperparameters, and assess the impact of instruction dataset size, showing that a larger number of parameters and a larger training dataset do not necessarily result in better performance for PEFT. Lastly, we analyze peak memory usage and trainable parameters to show the efficiency of PEFT. This work provides a comprehensive exploration of PEFT on APR and suggests potentially promising directions for extension to other software engineering downstream tasks. APR-INSTRUCTION, PEFT weights, and the fine-tuning code are publicly available as open-source resources.

cs.SE↗

arXiv2024

Relative Time-of-Flight Measurement in an Ultrafast Electron Microscope

Efforts to push the spatiotemporal imaging-resolution limits of femtosecond (fs) laser-driven ultrafast electron microscopes (UEMs) to the combined angstrom-fs range will benefit from stable sources capable of generating high bunch charges. Recent demonstration of unconventional off-axis photoemitting geometries are promising, but connections to the observed onset of structural dynamics are yet to be established. Here we use the in-situ photoexcitation of coherent phonons to quantify the relative time-of-flight (r-TOF) of photoelectron packets generated from the Ni Wehnelt aperture and from a Ta cathode set-back from the aperture plane. We further support the UEM experiments with particle-tracing simulations of the precise electron-gun architecture and photoemitting geometries. In this way, we measure discernable shifts in electron-packet TOF of tens of picoseconds for the two photoemitting surfaces. These shifts arise from the impact the Wehnelt-aperture off-axis orientation has on the electron-momentum distribution, which modifies both the collection efficiency and the temporal-packet distribution relative to on-axis emission. Future needs are identified; we expect this and other developments in UEM electron-gun configuration to expand the range of materials phenomena that can be directly imaged on scales commensurate with fundamental structural dynamics.

cond-mat.mtrl-sci↗

arXiv2022

Learning to Evaluate Performance of Multi-modal Semantic Localization

Semantic localization (SeLo) refers to the task of obtaining the most relevant locations in large-scale remote sensing (RS) images using semantic information such as text. As an emerging task based on cross-modal retrieval, SeLo achieves semantic-level retrieval with only caption-level annotation, which demonstrates its great potential in unifying downstream tasks. Although SeLo has been carried out successively, but there is currently no work has systematically explores and analyzes this urgent direction. In this paper, we thoroughly study this field and provide a complete benchmark in terms of metrics and testdata to advance the SeLo task. Firstly, based on the characteristics of this task, we propose multiple discriminative evaluation metrics to quantify the performance of the SeLo task. The devised significant area proportion, attention shift distance, and discrete attention distance are utilized to evaluate the generated SeLo map from pixel-level and region-level. Next, to provide standard evaluation data for the SeLo task, we contribute a diverse, multi-semantic, multi-objective Semantic Localization Testset (AIR-SLT). AIR-SLT consists of 22 large-scale RS images and 59 test cases with different semantics, which aims to provide a comprehensive evaluations for retrieval models. Finally, we analyze the SeLo performance of RS cross-modal retrieval models in detail, explore the impact of different variables on this task, and provide a complete benchmark for the SeLo task. We have also established a new paradigm for RS referring expression comprehension, and demonstrated the great advantage of SeLo in semantics through combining it with tasks such as detection and road extraction. The proposed evaluation metrics, semantic localization testsets, and corresponding scripts have been open to access at github.com/xiaoyuan1996/SemanticLocalizationMetrics .

cs.CV↗

arXiv2022

Strolling in Room-Scale VR: Hex-Core-MK1 Omnidirectional Treadmill

The natural locomotion interface is critical to the development of many VR applications. For household VR applications, there are two basic requirements: natural immersive experience and minimized space occupation. The existing locomotion strategies generally do not simultaneously satisfy these two requirements well. This paper presents a novel omnidirectional treadmill (ODT) system, named Hex-Core-MK1 (HCMK1). By implementing two kinds of mirror symmetrical spiral rollers to generate the omnidirectional velocity field, this proposed system is capable of providing real walking experiences with a full-degree of freedom in an area as small as 1.76 m^2, while delivering great advantages over several existing ODT systems in terms of weight, volume, latency and dynamic performance. Compared with the sizes of Infinadeck and HCP, the two best motor-driven ODTs so far, the 8 cm height of HCMK1 is only 20% of Infinadeck and 50% of HCP. In addition, HCMK1 is a lightweight device weighing only 110 kg, which provides possibilities of further expanding VR scenarios, such as terrain simulation. The latency of HCMK1 is only 23ms. The experiments show that HCMK1 can deliver on a starting acceleration of 16.00 m/s^2 and a braking acceleration of 30.00 m/s^2.

cs.HC↗