arXiv subjects

Yucheng Jin

Publications and source records attributed to Yucheng Jin.

18 recordsLinked to original sources

arXiv2026

Visible but Not Yet Curatable: Characterizing the Curatability of Compact and Derived Open LLM Artifacts

Open Large Language Model (LLM) research increasingly produces compact and derived artifacts, such as adapters, quantized checkpoints, merged models, and distilled variants, that are distributed across papers, model hubs, model cards, code repositories, and release statements. Although these artifacts are publicly visible, digital libraries often lack sufficient evidence to identify, preserve, and cite them as coherent scholarly objects. We introduce a framework that conceptualizes curatability as a record-level property of distributed scholarly records and operationalizes it through four evidence dimensions: artifact identity, scholarly linkage, upstream evidence, and release assets. Guided by this framework, we conduct the first collection-scale characterization of open LLM curatability using a May 2026 snapshot of 191,375 public Hugging Face repositories and a core corpus of 2,214 scholarly papers. Our results reveal a pronounced visibility-to-curatability funnel. While 90.7% of paper records contain at least one useful curation signal, only 18.1% combine usable upstream evidence with concrete release evidence, and only 6.1% provide sufficiently coordinated evidence to support high-curatability records. Based on these findings, we derive a minimal seven-field curatable record and complementary responsibilities for model hubs, scholarly indexes, and digital libraries, providing practical guidance for improving the preservation and bibliographic control of open LLM artifacts.

cs.HC↗

arXiv2026

How Much Regularization Survives Averaging? Update Masking in Federated Learning

Federated learning on non-IID data seeks flat minima to generalize across clients, and existing methods borrow sharpness-aware minimization from centralized training. There is a second way to reach flat minima, in which the regularization comes for free from noise added to the parameter updates, and it has never been carried over to the federated setting as an implicit regularizer. We show the reason. Masking charges the optimizer for moving in sharp directions. We prove that when each client draws its own mask, federated averaging weakens that charge by exactly the cohort size, and that giving every client the same mask brings it back by a factor equal to the inverse gradient diversity of the cohort. In our experiment setting on CIFAR-10, that factor is 1.19 out of a possible 10. Turning off minibatch sampling raises it to 8.96, while changing data heterogeneity a thousandfold leaves it between 1.17 and 1.50. The configurations keeping the regularization train far too poorly to use.

cs.LG↗

arXiv2026

CrystalREPA: Transferring Physical Priors from Universal MLIPs to Crystal Generative Models

Crystal generative models mainly learn what stable crystals look like, with little explicit supervision for what makes them stable. We reveal a substantial representation gap between state-of-the-art crystal generative models and pretrained universal machine learning interatomic potentials (MLIPs) via energy probing, and show this gap can be closed by a simple training-time alignment. We propose Crystal REPresentation Alignment (CrystalREPA), a plug-and-play framework that aligns the atom-wise hidden states of generative encoders with frozen MLIP representations through an element-aware contrastive objective, transferring stability-aware atomistic priors with marginal training overhead and no additional inference cost. Across three generative frameworks, ten MLIP teachers, and two benchmark datasets, CrystalREPA consistently improves the thermodynamic stability, structural validity, and structural fidelity of generated crystals. Equally important, we find that an MLIP's transfer effectiveness is poorly predicted by its accuracy on standard leaderboards (e.g., Matbench Discovery) but strongly predicted by the distinguishability of its atom-wise representation space, yielding a practical, accuracy-independent criterion for selecting MLIP teachers for generative transfer.

cond-mat.mtrl-sci↗

arXiv2026

How Personal Characteristics Shape User Exploration of Diverse Movie Recommendations with a LLM-Based Multi-Agent System

Diversity is an important evaluation criterion for recommender systems beyond accuracy, yet users differ in their willingness to engage with novel and diverse content. In this work, we investigate how a Large Language Model (LLM)-based multi-agent system supports users' exploration of diverse recommendations, and how individual characteristics shape user experiences. We conducted a between-subjects user study (N = 100) comparing a single-agent system (baseline) with a multi-agent system for movie recommendations. We measured Perceived Accuracy, diversity, novelty, and overall rating, and examined the influence of personal characteristics, including personality traits, demographics, GenAI recommendation experience, and GenAI skepticism. Results show that the multi-agent system significantly increases Perceived Novelty and Shannon Diversity. Conscientiousness is positively associated with Perceived Accuracy and diversity, whereas extraversion is negatively associated with Perceived Diversity. Prior experience with GenAI-based recommendations is positively associated with Shannon Diversity, while skepticism toward GenAI is negatively associated with it. We also observe significant interaction effects between system design and user characteristics. These findings highlight the importance of personality-aware conversational recommender systems and caution against one-size-fits-all multi-agent designs.

cs.HC↗

arXiv2026

Towards Faithful Reasoning in Comics for Small MLLMs

Comic understanding presents a significant challenge for Multimodal Large Language Models (MLLMs), as the intended meaning of a comic often emerges from the joint interpretation of visual, textual, and social cues. This naturally motivates Chain-of-Thought (CoT) prompting, since explicit intermediate reasoning appears promising for integrating such heterogeneous signals. However, existing CoT methods are poorly matched to this structure: they tend to force interpretation into a single reasoning path before multiple cues have been jointly considered, often degrading performance, especially for small MLLMs. Our key idea is to explicitly preserve multi-cue interpretation during supervision construction, rather than collapsing comic understanding into a single reasoning chain. To this end, we propose a two-stage framework for faithful comic reasoning in small MLLMs. First, we introduce MoCoT, a modular supervision construction framework that preserves multi-cue interpretation and turns it into more faithful supervision. Second, we propose VERA, a structured reward mechanism that turns such supervision into faithful reasoning behavior by aligning optimization with both reasoning faithfulness and answer correctness. Extensive experiments on five benchmarks spanning comic understanding and broader humor-centric and abstract visual reasoning tasks demonstrate that our framework achieves strong results in the $\leq$ 4B regime, surpasses several 7B baselines, improves four small MLLMs by an average of $\mathbf{12.1%}$ as a plug-in, and consistently enhances reasoning faithfulness while preserving inference efficiency.

cs.CV↗

arXiv2025

Colin: A Multimodal Human-AI Co-Creation Storytelling System To Support Children's Multi-Level Narrative Skills

Children develop narrative skills by understanding and actively building connections between elements, image text matching, and consequences. However, it is challenging for children to clearly grasp these multi level links only through explanations of text or the facilitator's speech. To address this, we developed Colin, an interactive storytelling tool that supports children's multi level narrative skills through both voice and visual modalities. In the generation stage, Colin supports the facilitator to define and review the generated text and image content freely. In the understanding stage, a question feedback model helps children understand multi level connections while co creating stories with Colin. In the building phase, Colin actively encourages children to create connections between elements through drawing and speaking. A user study with 20 participants evaluated Colin by measuring children's engagement, understanding of cause and effect relationships, and the quality of their new story creations. Our results demonstrate that Colin significantly enhances the development of children's narrative skills across multiple levels.

cs.HC↗

arXiv2025

Hyperspectral Dual-Comb Compressive Imaging for Minimally-Invasive Video-Rate Endomicroscopy

Endoscopic imaging is essential for real-time visualization of internal organs, yet conventional systems remain bulky, complex, and expensive due to their reliance on large, multi-element optical components. This limits their accessibility to delicate or constrained anatomical regions. Achieving real-time, high-resolution endomicroscopy using compact, low-cost hardware at the hundred-micron scale remains an unsolved challenge. Optical fibers offer a promising route toward miniaturization by providing sub-millimeter-scale imaging channels; however, existing fiber-based methods typically rely on raster scanning or multicore bundles, which limit the resolution and imaging speed. In this work, we overcome these limitations by integrating dual-comb interferometry with compressive ghost imaging and advanced computational reconstruction. Our technique, hyperspectral dual-comb compressive imaging, utilizes optical frequency combs to generate wavelength-multiplexed speckle patterns that are delivered through a single-core fiber and detected by a single-pixel photodetector. This parallel speckle illumination and detection enable snapshot compression and acquisition of image information using zero-dimensional hardware, completely eliminating the need for both spatial and spectral scanning. To decode these highly compressed signals, we develop a transformer-based deep learning model capable of rapid, high-fidelity image reconstruction at extremely low sampling ratios. This approach significantly outperforms classical ghost imaging methods in both speed and accuracy, achieving video-rate imaging with a dramatically simplified optical front-end. Our results represent a major advance toward minimally invasive, cost-effective endomicroscopy and provide a generalizable platform for optical sensing in applications where hardware constraints are critical.

physics.optics↗

arXiv2025

STAR-Pose: Efficient Low-Resolution Video Human Pose Estimation via Spatial-Temporal Adaptive Super-Resolution

Human pose estimation in low-resolution videos presents a fundamental challenge in computer vision. Conventional methods either assume high-quality inputs or employ computationally expensive cascaded processing, which limits their deployment in resource-constrained environments. We propose STAR-Pose, a spatial-temporal adaptive super-resolution framework specifically designed for video-based human pose estimation. Our method features a novel spatial-temporal Transformer with LeakyReLU-modified linear attention, which efficiently captures long-range temporal dependencies. Moreover, it is complemented by an adaptive fusion module that integrates parallel CNN branch for local texture enhancement. We also design a pose-aware compound loss to achieve task-oriented super-resolution. This loss guides the network to reconstruct structural features that are most beneficial for keypoint localization, rather than optimizing purely for visual quality. Extensive experiments on several mainstream video HPE datasets demonstrate that STAR-Pose outperforms existing approaches. It achieves up to 5.2% mAP improvement under extremely low-resolution (64x48) conditions while delivering 2.8x to 4.4x faster inference than cascaded approaches.

cs.CV↗

arXiv2024

Navigating User Experience of ChatGPT-based Conversational Recommender Systems: The Effects of Prompt Guidance and Recommendation Domain

Conversational recommender systems (CRS) enable users to articulate their preferences and provide feedback through natural language. With the advent of large language models (LLMs), the potential to enhance user engagement with CRS and augment the recommendation process with LLM-generated content has received increasing attention. However, the efficacy of LLM-powered CRS is contingent upon the use of prompts, and the subjective perception of recommendation quality can differ across various recommendation domains. Therefore, we have developed a ChatGPT-based CRS to investigate the impact of these two factors, prompt guidance (PG) and recommendation domain (RD), on the overall user experience of the system. We conducted an online empirical study (N = 100) by employing a mixed-method approach that utilized a between-subjects design for the variable of PG (with vs. without) and a within-subjects design for RD (book recommendations vs. job recommendations). The findings reveal that PG can substantially enhance the system's explainability, adaptability, perceived ease of use, and transparency. Moreover, users are inclined to perceive a greater sense of novelty and demonstrate a higher propensity to engage with and try recommended items in the context of book recommendations as opposed to job recommendations. Furthermore, the influence of PG on certain user experience metrics and interactive behaviors appears to be modulated by the recommendation domain, as evidenced by the interaction effects between the two examined factors. This work contributes to the user-centered evaluation of ChatGPT-based CRS by investigating two prominent factors and offers practical design guidance.

cs.HC↗

arXiv2024

Understanding Human-AI Collaboration in Music Therapy Through Co-Design with Therapists

The rapid development of musical AI technologies has expanded the creative potential of various musical activities, ranging from music style transformation to music generation. However, little research has investigated how musical AIs can support music therapists, who urgently need new technology support. This study used a mixed method, including semi-structured interviews and a participatory design approach. By collaborating with music therapists, we explored design opportunities for musical AIs in music therapy. We presented the co-design outcomes involving the integration of musical AIs into a music therapy process, which was developed from a theoretical framework rooted in emotion-focused therapy. After that, we concluded the benefits and concerns surrounding music AIs from the perspective of music therapists. Based on our findings, we discussed the opportunities and design implications for applying musical AIs to music therapy. Our work offers valuable insights for developing human-AI collaborative music systems in therapy involving complex procedures and specific requirements.

cs.HC↗

arXiv2024

Beyond the Known: Novel Class Discovery for Open-world Graph Learning

Node classification on graphs is of great importance in many applications. Due to the limited labeling capability and evolution in real-world open scenarios, novel classes can emerge on unlabeled testing nodes. However, little attention has been paid to novel class discovery on graphs. Discovering novel classes is challenging as novel and known class nodes are correlated by edges, which makes their representations indistinguishable when applying message passing GNNs. Furthermore, the novel classes lack labeling information to guide the learning process. In this paper, we propose a novel method Open-world gRAph neuraL network (ORAL) to tackle these challenges. ORAL first detects correlations between classes through semi-supervised prototypical learning. Inter-class correlations are subsequently eliminated by the prototypical attention network, leading to distinctive representations for different classes. Furthermore, to fully explore multi-scale graph features for alleviating label deficiencies, ORAL generates pseudo-labels by aligning and ensembling label estimations from multiple stacked prototypical attention networks. Extensive experiments on several benchmark datasets show the effectiveness of our proposed method.

cs.LG↗

arXiv2024

Toward parallel intelligence: an interdisciplinary solution for complex systems

The growing complexity of real-world systems necessitates interdisciplinary solutions to confront myriad challenges in modeling, analysis, management, and control. To meet these demands, the parallel systems method rooted in Artificial systems, Computational experiments, and Parallel execution (ACP) approach has been developed. The method cultivates a cycle, termed parallel intelligence, which iteratively creates data, acquires knowledge, and refines the actual system. Over the past two decades, the parallel systems method has continuously woven advanced knowledge and technologies from various disciplines, offering versatile interdisciplinary solutions for complex systems across diverse fields. This review explores the origins and fundamental concepts of the parallel systems method, showcasing its accomplishments as a diverse array of parallel technologies and applications, while also prognosticating potential challenges. We posit that this method will considerably augment sustainable development while enhancing interdisciplinary communication and cooperation.

cs.DC↗

arXiv2024

Exploring the Design of Generative AI in Supporting Music-based Reminiscence for Older Adults

Music-based reminiscence has the potential to positively impact the psychological well-being of older adults. However, the aging process and physiological changes, such as memory decline and limited verbal communication, may impede the ability of older adults to recall their memories and life experiences. Given the advanced capabilities of generative artificial intelligence (AI) systems, such as generated conversations and images, and their potential to facilitate the reminiscing process, this study aims to explore the design of generative AI to support music-based reminiscence in older adults. This study follows a user-centered design approach incorporating various stages, including detailed interviews with two social workers and two design workshops (involving ten older adults). Our work contributes to an in-depth understanding of older adults' attitudes toward utilizing generative AI for supporting music-based reminiscence and identifies concrete design considerations for the future design of generative AI to enhance the reminiscence experience of older adults.

cs.HC↗

arXiv2023

NFT.mine: An xDeepFM-based Recommender System for Non-fungible Token (NFT) Buyers

Non-fungible token (NFT) is a tradable unit of data stored on the blockchain which can be associated with some digital asset as a certification of ownership. The past several years have witnessed the exponential growth of the NFT market. In 2021, the NFT market reached its peak with more than $40 billion trades. Despite the booming NFT market, most NFT-related studies focus on its technical aspect, such as standards, protocols, and security, while our study aims at developing a pioneering recommender system for NFT buyers. In this paper, we introduce an extreme deep factorization machine (xDeepFM)-based recommender system, NFT.mine, which achieves real-time data collection, data cleaning, feature extraction, training, and inference. We used data from OpenSea, the most influential NFT trading platform, to testify the performance of NFT.mine. As a result, experiments showed that compared to traditional models such as logistic regression, naive Bayes, random forest, etc., NFT.mine outperforms them with higher AUC and lower cross entropy loss and outputs personalized recommendations for NFT buyers.

cs.CR↗

arXiv2023

TIGER: Temporal Interaction Graph Embedding with Restarts

Temporal interaction graphs (TIGs), consisting of sequences of timestamped interaction events, are prevalent in fields like e-commerce and social networks. To better learn dynamic node embeddings that vary over time, researchers have proposed a series of temporal graph neural networks for TIGs. However, due to the entangled temporal and structural dependencies, existing methods have to process the sequence of events chronologically and consecutively to ensure node representations are up-to-date. This prevents existing models from parallelization and reduces their flexibility in industrial applications. To tackle the above challenge, in this paper, we propose TIGER, a TIG embedding model that can restart at any timestamp. We introduce a restarter module that generates surrogate representations acting as the warm initialization of node representations. By restarting from multiple timestamps simultaneously, we divide the sequence into multiple chunks and naturally enable the parallelization of the model. Moreover, in contrast to previous models that utilize a single memory unit, we introduce a dual memory module to better exploit neighborhood information and alleviate the staleness problem. Extensive experiments on four public datasets and one industrial dataset are conducted, and the results verify both the effectiveness and the efficiency of our work.

cs.LG↗

arXiv2022

Reinforcement Learning for Resilient Power Grids

Traditional power grid systems have become obsolete under more frequent and extreme natural disasters. Reinforcement learning (RL) has been a promising solution for resilience given its successful history of power grid control. However, most power grid simulators and RL interfaces do not support simulation of power grid under large-scale blackouts or when the network is divided into sub-networks. In this study, we proposed an updated power grid simulator built on Grid2Op, an existing simulator and RL interface, and experimented on limiting the action and observation spaces of Grid2Op. By testing with DDQN and SliceRDQN algorithms, we found that reduced action spaces significantly improve training performance and efficiency. In addition, we investigated a low-rank neural network regularization method for deep Q-learning, one of the most widely used RL algorithms, in this power grid control scenario. As a result, the experiment demonstrated that in the power grid simulation environment, adopting this method will significantly increase the performance of RL agents.

cs.LG↗

arXiv2022

Identifying Exoplanets with Machine Learning Methods: A Preliminary Study

The discovery of habitable exoplanets has long been a heated topic in astronomy. Traditional methods for exoplanet identification include the wobble method, direct imaging, gravitational microlensing, etc., which not only require a considerable investment of manpower, time, and money, but also are limited by the performance of astronomical telescopes. In this study, we proposed the idea of using machine learning methods to identify exoplanets. We used the Kepler dataset collected by NASA from the Kepler Space Observatory to conduct supervised learning, which predicts the existence of exoplanet candidates as a three-categorical classification task, using decision tree, random forest, naïve Bayes, and neural network; we used another NASA dataset consisted of the confirmed exoplanets data to conduct unsupervised learning, which divides the confirmed exoplanets into different clusters, using k-means clustering. As a result, our models achieved accuracies of 99.06%, 92.11%, 88.50%, and 99.79%, respectively, in the supervised learning task and successfully obtained reasonable clusters in the unsupervised learning task.

astro-ph.EP↗

arXiv2022

Impacts of Personal Characteristics on User Trust in Conversational Recommender Systems

Conversational recommender systems (CRSs) imitate human advisors to assist users in finding items through conversations and have recently gained increasing attention in domains such as media and e-commerce. Like in human communication, building trust in human-agent communication is essential given its significant influence on user behavior. However, inspiring user trust in CRSs with a "one-size-fits-all" design is difficult, as individual users may have their own expectations for conversational interactions (e.g., who, user or system, takes the initiative), which are potentially related to their personal characteristics. In this study, we investigated the impacts of three personal characteristics, namely personality traits, trust propensity, and domain knowledge, on user trust in two types of text-based CRSs, i.e., user-initiative and mixed-initiative. Our between-subjects user study (N=148) revealed that users' trust propensity and domain knowledge positively influenced their trust in CRSs, and that users with high conscientiousness tended to trust the mixed-initiative system.

cs.HC↗