arXiv subjects

Yining Zhao

Publications and source records attributed to Yining Zhao.

13 recordsLinked to original sources

arXiv2026

Can Large Language Models Forecast What Researchers Study Next?

Large language models increasingly generate research ideas, yet judging their novelty or feasibility at generation time does not establish whether they anticipate subsequent work. We introduce IdeaForecastBench to evaluate research idea forecasting. Given a community's literature up to a cutoff, a system produces up to five ranked ideas, which are evaluated against later papers. The benchmark comprises 624 rolling episodes across 52 topics, with a fixed retrieve-then-judge protocol and separately reported results from two judges. We compare five history-compression strategies across GPT-4.1, Qwen2.5-7B/14B, and Qwen3.5-9B, together with a learned Mode-Decomposition Forecaster (MDF). Under the primary GPT-4.1-mini judge, Summary improves on Direct in Hit@5 and Precision@5 across all four backbones. Qwen2.5 scores above GPT-4.1, whereas Qwen3.5 scores below it. An outcome-blind assessment finds that Qwen2.5 produces broader forecasts, but does not identify how much breadth contributes to its advantage. Threshold and judge diagnostics further clarify the limits of interpreting realization as precise anticipation. IdeaForecastBench provides a common task for studying which research ideas a community subsequently pursues and how reliably this outcome can be measured.

cs.CL↗

arXiv2026

TARE: Tail Aware Evaluation of HPC Job Runtime Prediction

Runtime estimates affect reservation quality, backfilling opportunities, and queue delay in HPC schedulers. Under heavy tailed workloads, however, averaging over jobs can misrepresent scheduling impact because a small fraction of jobs dominates resource usage. This paper presents an empirical evaluation methodology for HPC job runtime prediction that focuses on the tail, combining GeoAccuracy weighted by resource usage with decile and split analyses. Using production traces from NREL Eagle and ALCF Mira/Intrepid, we compare XGBoost and Last2 against the user provided walltime estimate at submission (UserReq). Across all three datasets, evaluation focused on the tail changes the offline conclusion: MeanAccuracy keeps the methods relatively close, whereas GeoAccuracy reveals clearer separation and makes UserReq's strength in the upper tail visible. In the top decile, UserReq achieves the highest GeoAccuracy and lowest underestimation rate on all three datasets, and this pattern remains stable across rolling splits. We then translate this signal into a simple hybrid scheduling policy that keeps XGBoost for most jobs and routes the top decile by proxy_cost at submission to UserReq. Online replay on four production queues reduces mean wait time by up to 8% and increases backfilled jobs by 50% to 115%. These results show that offline evaluation focused on the tail better characterizes prediction quality relevant to scheduling and informs scheduling policy design.

cs.DC↗

arXiv2026

Building Social World Models with Large Language Models

Understanding and predicting how social beliefs evolve in response to events -- from policy changes to scientific breakthroughs -- remains a fundamental challenge in social science. Given LLMs' commonsense knowledge and social intelligence, we ask: Can LLMs model the dynamics of social beliefs following social events? In this work, we introduce the concept of the Social World Model (SWM), a general framework designed to capture how social beliefs evolve in response to major events. SWM learns state-transition functions for social beliefs by mining temporal patterns in social data and optimizing the evidence lower bound, without the need for explicit human annotations linking events to belief shifts, or for expensive census data. To evaluate SWM, we introduce a benchmark, SWM-bench, derived from real-world prediction markets, specifically Kalshi and Polymarket. SWM-bench includes over 12k data points for social belief prediction tasks spanning diverse domains such as politics, finance, and cryptocurrency. Our experimental results show that SWM significantly outperforms time-series foundation models, achieving state-of-the-art results on Kalshi data and demonstrating competitive performance on Polymarket data, while offering interpretable insights into the underlying mechanisms of social belief dynamics.

cs.SI↗

arXiv2026

Auto-Dreamer: Learning Offline Memory Consolidation for Language Agents

Language agents increasingly operate over streams of related tasks, yet existing memory systems struggle to convert accumulated experience into reusable knowledge. Retrieval-augmented and structured memory methods record per-session observations effectively, but often couple acquisition and consolidation into a single online process, leaving the agent without a global view across sessions to discover recurring patterns, abstract shared procedures, or prune redundant entries. Inspired by complementary learning systems theory, we propose Auto-Dreamer, a learned offline consolidator for language-agent memory. Auto-Dreamer decouples fast per-session memory acquisition from slow cross-session consolidation. Given a selected working region of a typed memory bank, the consolidator treats the region as read-only evidence, performs bounded tool-use to inspect entries and provenance-linked source trajectories, and synthesizes a fresh compact replacement set that abstracts across sessions and supersedes the original region. We train Auto-Dreamer via GRPO, using end-to-end agent performance as the reward signal to learn how to consolidate memories acquired through fast online experience. Trained on ScienceWorld trajectories alone, Auto-Dreamer outperforms fixed, RL-trained, and prompted memory baselines on ScienceWorld by 7 points while using an active memory bank 12$\times$ smaller than the strongest baseline, and continues to lead on held-out ALFWorld and WebArena without retraining -- using 6$\times$ less memory than the strongest baseline on ALFWorld.

cs.CL↗

arXiv2026

CTM-AI: A Blueprint for General AI Inspired by a Model of Consciousness

Despite remarkable advances, today's AI systems remain narrow in scope, falling short of the flexible, adaptive, and multisensory intelligence that characterizes human capabilities. This gap has fueled longstanding debates about whether AI might one day achieve human-like generality or even consciousness, and whether theories of consciousness can inspire new architectures for AI. This paper presents an early blueprint for implementing a general AI system, CTM-AI, combining the Conscious Turing Machine (CTM), a formal machine model of consciousness, with today's foundation models. CTM-AI contains an enormous number of powerful processors ranging from specialized experts (e.g., vision-language models and APIs) to unspecialized general-purpose learners poised to develop their own expertise. Crucially, for whatever problem must be dealt with, information from many processors is selected, integrated, and exchanged appropriately to solve the task. CTM-AI achieves state-of-the-art accuracy on MUStARD (72.28) and UR-FUNNY (72.13), outperforming multimodal and multi-agent frameworks. On tool-using and agentic tasks, CTM-AI achieves 10+ points of improvement on StableToolBench and WebArena-Lite. Overall, CTM-AI offers a principled, testable blueprint for general AI inspired by a model of consciousness.

q-bio.NC↗

arXiv2026

Meet Dynamic Individual Preferences: Resolving Conflicting Human Value with Paired Fine-Tuning

Recent advances in large language models (LLMs) have significantly improved the alignment of models with general human preferences. However, a major challenge remains in adapting LLMs to individual preferences, which are not only diverse but also dynamic. In this paper, we introduce a novel framework, Preference-Paired Fine-Tuning (PFT), designed to align models with contradictory and evolving individual preferences. We present a new dataset, Value Conflict Dilemma (VCD), which includes scenarios that involve conflicting human preferences, facilitating the evaluation of our approach. Our experiments demonstrate that PFT outperforms single-preference training methods, achieving up to 96.6% accuracy in multi-choice classification tasks and the highest open-ended generation score of 8.69. PFT also shows significant improvements over DPO, SFT and some traditional training methods, especially when handling conflicting preferences. Additionally, with limited user history data, models can inferring preference vector rapidly, achieving a 44.76% improvement in user-specific preference alignment in comparison to single-preference models.

cs.CL↗

arXiv2025

A User-customized and Untethered Electro-haptic Device for Immersive Human-Machine Interaction

Haptic feedback is essential for human-machine interaction, as it bridges physical and digital experiences and enables immersive engagement with virtual environments. However, current haptic devices are frequently tethered, lack portability and flexibility. They also have limited ability to deliver fine-grained, multi-dimensional feedback. To address these challenges, we present a flexible, ultra-thin, and user-customized electro-haptic device fabricated with soft materials and printable liquid metal ink. Its highly integrated and lightweight design minimizes interference with natural hand movements while maintaining reliable skin contact. By delivering finely controlled electrical stimulation through 15 electrodes, it can evoke a wide range of tactile sensations that cover diverse interaction scenarios. Our user study demonstrates that the device is comfortable to wear and capable of generating tunable, precise electro-haptic feedback, thereby significantly enhancing immersion and realism in human-machine interactions.

cs.HC↗

arXiv2025

Sotopia-RL: Reward Design for Social Intelligence

Social intelligence has become a critical capability for large language models (LLMs), enabling them to engage effectively in real-world social tasks such as collaboration and negotiation. Reinforcement learning (RL) is a natural fit for training socially intelligent agents because it allows models to learn sophisticated strategies directly through social interactions without requiring human annotations. However, there are two unique parts about social intelligence tasks: (1) the quality of individual utterances in social interactions is not strictly related to final success; (2) social interactions require multi-dimensional rubrics for success. Therefore, we argue that it is necessary to design rewards for building utterance-level multi-dimensional reward models to facilitate RL training for social intelligence tasks. To address these challenges, we propose Sotopia-RL, a novel framework that refines coarse episode-level feedback into utterance-level, multi-dimensional rewards. Utterance-level credit assignment attributes outcomes to individual utterances, while multi-dimensional rewards capture the full richness of social interactions and reduce reward hacking. Experiments in Sotopia, an open-ended social learning environment, demonstrate that Sotopia-RL achieves state-of-the-art social goal completion scores (7.17 on Sotopia-hard and 8.31 on Sotopia-full), significantly outperforming existing approaches. Ablation studies confirm the necessity of both utterance-level credit assignment and multi-dimensional reward design for RL training.

cs.CL↗

arXiv2025

How Cohesive Are Community Search Results on Online Social Networks?: An Experimental Evaluation

Recently, numerous community search methods for large graphs have been proposed, at the core of which is defining and measuring cohesion. This paper experimentally evaluates the effectiveness of these community search algorithms w.r.t. cohesiveness in the context of online social networks. Social communities are formed and developed under the influence of group cohesion theory, which has been extensively studied in social psychology. However, current generic methods typically measure cohesiveness using structural or attribute-based approaches and overlook domain-specific concepts such as group cohesion. We introduce five novel psychology-informed cohesiveness measures, based on the concept of group cohesion from social psychology, and propose a novel framework called CHASE for evaluating eight representative community search algorithms w.r.t. these measures on online social networks. Our analysis reveals that there is no clear correlation between structural and psychological cohesiveness, and no algorithm effectively identifies psychologically cohesive communities in online social networks. This study provides new insights that could guide the development of future community search methods.

cs.IR↗

arXiv2025

DualPrompt-MedCap: A Dual-Prompt Enhanced Approach for Medical Image Captioning

Medical image captioning via vision-language models has shown promising potential for clinical diagnosis assistance. However, generating contextually relevant descriptions with accurate modality recognition remains challenging. We present DualPrompt-MedCap, a novel dual-prompt enhancement framework that augments Large Vision-Language Models (LVLMs) through two specialized components: (1) a modality-aware prompt derived from a semi-supervised classification model pretrained on medical question-answer pairs, and (2) a question-guided prompt leveraging biomedical language model embeddings. To address the lack of captioning ground truth, we also propose an evaluation framework that jointly considers spatial-semantic relevance and medical narrative quality. Experiments on multiple medical datasets demonstrate that DualPrompt-MedCap outperforms the baseline BLIP-3 by achieving a 22% improvement in modality recognition accuracy while generating more comprehensive and question-aligned descriptions. Our method enables the generation of clinically accurate reports that can serve as medical experts' prior knowledge and automatic annotations for downstream vision-language tasks.

cs.CV↗

arXiv2025

Implicit U-KAN2.0: Dynamic, Efficient and Interpretable Medical Image Segmentation

Image segmentation is a fundamental task in both image analysis and medical applications. State-of-the-art methods predominantly rely on encoder-decoder architectures with a U-shaped design, commonly referred to as U-Net. Recent advancements integrating transformers and MLPs improve performance but still face key limitations, such as poor interpretability, difficulty handling intrinsic noise, and constrained expressiveness due to discrete layer structures, often lacking a solid theoretical foundation.In this work, we introduce Implicit U-KAN 2.0, a novel U-Net variant that adopts a two-phase encoder-decoder structure. In the SONO phase, we use a second-order neural ordinary differential equation (NODEs), called the SONO block, for a more efficient, expressive, and theoretically grounded modeling approach. In the SONO-MultiKAN phase, we integrate the second-order NODEs and MultiKAN layer as the core computational block to enhance interpretability and representation power. Our contributions are threefold. First, U-KAN 2.0 is an implicit deep neural network incorporating MultiKAN and second order NODEs, improving interpretability and performance while reducing computational costs. Second, we provide a theoretical analysis demonstrating that the approximation ability of the MultiKAN block is independent of the input dimension. Third, we conduct extensive experiments on a variety of 2D and a single 3D dataset, demonstrating that our model consistently outperforms existing segmentation networks. Project Website: https://math-ml-x.github.io/IUKAN2/

eess.IV↗

arXiv2022

3D Room Layout Estimation from a Cubemap of Panorama Image via Deep Manhattan Hough Transform

Significant geometric structures can be compactly described by global wireframes in the estimation of 3D room layout from a single panoramic image. Based on this observation, we present an alternative approach to estimate the walls in 3D space by modeling long-range geometric patterns in a learnable Hough Transform block. We transform the image feature from a cubemap tile to the Hough space of a Manhattan world and directly map the feature to the geometric output. The convolutional layers not only learn the local gradient-like line features, but also utilize the global information to successfully predict occluded walls with a simple network structure. Unlike most previous work, the predictions are performed individually on each cubemap tile, and then assembled to get the layout estimation. Experimental results show that we achieve comparable results with recent state-of-the-art in prediction accuracy and performance. Code is available at https://github.com/Starrah/DMH-Net.

cs.CV↗

arXiv2021

Quantifying human mixing patterns in Chinese provinces outside Hubei after the 2020 lockdown was lifted

Contact patterns play a key role in the spread of respiratory infectious diseases in human populations. During the COVID-19 pandemic the regular contact patterns of the population has been disrupted due to social distancing both imposed by the authorities and individual choices. Here we present the results of a contact survey conducted in Chinese provinces outside Hubei in March 2020, right after lockdowns were lifted. We then leveraged the estimated mixing patterns to calibrate a model of SARS-CoV-2 transmission, which was used to estimate different metrics of COVID-19 burden by age. Study participants reported 2.3 contacts per day (IQR: 1.0-3.0) and the mean per-contact duration was 7.0 hours (IQR: 1.0-10.0). No significant differences were observed between provinces, the number of recorded contacts did not show a clear-cut trend by age, and most of the recorded contacts occurred with family members (about 78%). Our findings suggest that, despite the lockdown was no longer in place at the time of the survey, people were still heavily limiting their contacts as compared to the pre-pandemic situation. Moreover, the obtained modeling results highlight the importance of considering age-contact patterns to estimate COVID-19 burden.

cs.SI↗