arXiv subjects

Yichen Gao

Publications and source records attributed to Yichen Gao.

8 recordsLinked to original sources

arXiv2026

SAILRec: Steering LLM Attention to Dual-Side Semantically Aligned Collaborative Embeddings for Recommendation

Recent LLM-based recommenders enhance language models with collaborative embeddings from user-item interactions, but making such embeddings available does not ensure their proper use during inference. Through a diagnostic attention analysis, we find that the utilization of collaborative embeddings is depth-dependent and alignment-sensitive, suggesting that LLMs need to balance their internal semantic knowledge with external collaborative knowledge. To address this issue, we propose SAILRec, an LLM-based recommender that improves this balance through dual-side semantic alignment and hierarchical attention steering. The former aligns item-side embeddings with item-text semantics and user-side embeddings with codebook-based semantic profiles, while the latter suppresses premature shallow-layer collaborative interference and strengthens collaborative evidence in deeper decision layers. Experiments on MovieLens-1M and Amazon-Book show that SAILRec consistently outperforms representative baselines, with ablation and masking analyses validating its key designs.

cs.IR↗

arXiv2026

Beyond Text Following: Repairable Arbitration Reversals in Audio-Language Models

Audio-language models (ALMs) often follow text that conflicts with audio, even when the audio evidence is clear. This raises a basic question: is the audio-supported answer unavailable, or is it represented but overridden by the conflicting text? We examine this question using a same-audio counterfactual that keeps the audio fixed, removes only the conflicting text, and measures the resulting shift in model preference. Across five ALMs and four conflict tasks, 64.1% of conflict samples show a sign flip: the same-audio branch prefers the audio-supported answer, whereas the joint branch prefers the text-supported answer. This pattern suggests that the relevant audio evidence is encoded but loses in arbitration. Activation patching further localizes the reversal to answer-position computation, and patching effects closely track output candidate-score differences (Spearman rho=0.93). Using this diagnostic, we propose Gated Audio Counterfactual Logit Correction (GACL), a training-free decoding rule that interpolates between joint and same-audio scores. Under a strict 5 pp faithfulness-drop budget, GACL improves nAUC by 17.8 points over the best contrastive baseline and transfers without retuning to vision-text arbitration (up to +40.5 pp).

cs.SD↗

arXiv2026

Agentic Artificial Intelligence in Finance: A Comprehensive Survey

The emergence of agentic artificial intelligence (AI) represents a fundamental transformation in financial markets, characterized by autonomous systems capable of reasoning, planning, and adaptive decision-making with minimal human intervention. This comprehensive survey synthesizes recent advances in agentic AI across multiple dimensions of financial operations, including system architecture, market applications, regulatory frameworks, and systemic implications. We examine how agentic AI differs from traditional algorithmic trading and generative AI through its capacity for goal-oriented autonomy, continuous learning, and multi-agent coordination. Our analysis shows that while agentic AI offers substantial potential for enhanced market efficiency, liquidity provision, and risk management, it also introduces novel challenges related to market stability, regulatory compliance, interpretability, and systemic risk. Through a systematic review of foundational research, technical architectures, market applications, and governance frameworks, this survey provides scholars and practitioners with a structured understanding of how agentic AI is reshaping financial markets and identifies critical research directions for ensuring that these systems enhance both operational efficiency and market resilience.

econ.EM↗

arXiv2026

An Illusion of Unlearning? Assessing Machine Unlearning Through Internal Representations

While numerous machine unlearning (MU) methods have recently been developed with promising results in erasing the influence of forgotten data, classes, or concepts, they are also highly vulnerable-for example, simple fine-tuning can inadvertently reintroduce erased concepts. In this paper, we address this contradiction by examining the internal representations of unlearned models, in contrast to prior work that focuses primarily on output-level behavior. Our analysis shows that many state-of-the-art MU methods appear successful mainly due to a misalignment between last-layer features and the classifier, a phenomenon we call feature-classifier misalignment. In fact, hidden features remain highly discriminative, and simple linear probing can recover near-original accuracy. Assuming neural collapse in the original model, we further demonstrate that adjusting only the classifier can achieve negligible forget accuracy while preserving retain accuracy, and we corroborate this with experiments using classifier-only fine-tuning. Motivated by these findings, we propose MU methods based on a class-mean features (CMF) classifier, which explicitly enforces alignment between features and classifiers. Experiments on standard benchmarks show that CMF-based unlearning reduces forgotten information in representations while maintaining high retain accuracy, highlighting the need for faithful representation-level evaluation of MU.

cs.LG↗

arXiv2025

Grating haptic perception through touchscreen: Sighted vs. Visually Impaired

Providing haptic feedback via smartphone touch screen may potentially offer blind people a capability to understand graphs. This study investigated the discrimination performance of haptic gratings in different frequencies, in both visually impaired (VI) and sighted (S) individuals. 6 VI participants and 10 S participants took part in two experiments designed to compare their ability to interpret grating images with a finger swiping across a smartphone touchscreen without vision. The swipe gesture activates phone vibration temporally synchronized with the black stripes. Their tasks were: (1) determining whether a grating pattern is presented on the touchscreen, (2) comparing two different grating frequencies and determining the wider one. Results demonstrated that the VI group exhibited superior tactile sensitivity compared to the S group, as evidenced by their significantly better performance in Experiment 1 (accuracy of 99.15\% vs. 84.5\%). Experiment 2 revealed that the peak performance of VI participants was approximately around 0.270 cycles per mm (83.3\% accuracy), a frequency very similar to Braille dot spacing, while S group peaked around 0.963 cycles per mm (70\% accuracy). The findings suggest that tactile stimulation coded with grating patterns could be potentially used to present interpretable graph for the visually impaired. Such an approach could offer a value to research in human-computer interaction and sensory adaptation.

cs.HC↗

arXiv2025

Bistability, Oscillations, and Multistability on Hycean Planets

Hycean planets are hypothetical exoplanets characterized by $H_2O$ oceans and $H_2$-rich atmospheres. These planets are high-priority targets for biosignature searches, as they combine abundant surface liquid water with easy-to-characterize $H_2$-rich atmospheres. Perhaps their most unusual climate feature is convective inhibition, which can dramatically alter a planet's temperature structure. However, so far hycean planets have mostly been investigated using 1D models that do not account for convective inhibition, and its effects are still poorly understood. This work develops pen-and-paper theory to analyze the effects of moist convective inhibition on hycean planets. The theory is tested and verified against a 1D radiative-convective model. We show that hycean planets near the onset of convective inhibition can exhibit either bistability or oscillations, due to the inhibition layer's trapping of heat and moisture. Meanwhile, hot hycean planets exhibit multistability, in which the inhibition layer and surface climate show multiple stable equilibria due to the lack of constraints on the water cycle inside the inhibition layer. The water cycle inside the inhibition layer is influenced by numerous processes that are challenging to resolve in 1D, including turbulent diffusion, convective overshoot and large-scale circulations. Our results demonstrate that hycean planets have unexpectedly rich climate dynamics. Meanwhile, previous claims about hycean planets should be treated with caution until confirmed with more self-consistent 1D and 3D models; this includes the claim that K2-18b might be habitable, and the proposal to infer $H_2O$ oceans on sub-Neptunes from JWST measurements of chemical species in their upper atmospheres.

astro-ph.EP↗

arXiv2024

Optimizing Global Quantum Communication via Satellite Constellations

In this paper, we investigate the optimization of global quantum communication through satellite constellations. We address the challenge of quantum key distribution (QKD) across vast distances and the limitations posed by terrestrial fiber-optic networks. Our research focuses on the configuration of satellite constellations to improve QKD between ground stations and the application of innovative orbital mechanics to reduce latency in quantum information transfer. We introduce a novel approach using quantum relay satellites in Molniya orbits, enhancing communication efficiency and coverage. The use of these high eccentricity orbits allows us to extend the operational presence of satellites over targeted hemispheres, thus maximizing the quantum network's reach. Our findings provide a strategic framework for deploying quantum satellites and relay systems to achieve a robust and efficient global quantum communication network.

quant-ph↗

arXiv2024

SeRTS: Self-Rewarding Tree Search for Biomedical Retrieval-Augmented Generation

Large Language Models (LLMs) have shown great potential in the biomedical domain with the advancement of retrieval-augmented generation (RAG). However, existing retrieval-augmented approaches face challenges in addressing diverse queries and documents, particularly for medical knowledge queries, resulting in sub-optimal performance. To address these limitations, we propose a novel plug-and-play LLM-based retrieval method called Self-Rewarding Tree Search (SeRTS) based on Monte Carlo Tree Search (MCTS) and a self-rewarding paradigm. By combining the reasoning capabilities of LLMs with the effectiveness of tree search, SeRTS boosts the zero-shot performance of retrieving high-quality and informative results for RAG. We further enhance retrieval performance by fine-tuning LLMs with Proximal Policy Optimization (PPO) objectives using the trajectories collected by SeRTS as feedback. Controlled experiments using the BioASQ-QA dataset with GPT-3.5-Turbo and LLama2-7b demonstrate that our method significantly improves the performance of the BM25 retriever and surpasses the strong baseline of self-reflection in both efficiency and scalability. Moreover, SeRTS generates higher-quality feedback for PPO training than self-reflection. Our proposed method effectively adapts LLMs to document retrieval tasks, enhancing their ability to retrieve highly relevant documents for RAG in the context of medical knowledge queries. This work presents a significant step forward in leveraging LLMs for accurate and comprehensive biomedical question answering.

cs.CL↗