arXiv subjects

Yu Shang

Publications and source records attributed to Yu Shang.

36 records · Page 2Linked to original sources

arXiv2025

AgentRecBench: Benchmarking LLM Agent-based Personalized Recommender Systems

The emergence of agentic recommender systems powered by Large Language Models (LLMs) represents a paradigm shift in personalized recommendations, leveraging LLMs' advanced reasoning and role-playing capabilities to enable autonomous, adaptive decision-making. Unlike traditional recommendation approaches, agentic recommender systems can dynamically gather and interpret user-item interactions from complex environments, generating robust recommendation strategies that generalize across diverse scenarios. However, the field currently lacks standardized evaluation protocols to systematically assess these methods. To address this critical gap, we propose: (1) an interactive textual recommendation simulator incorporating rich user and item metadata and three typical evaluation scenarios (classic, evolving-interest, and cold-start recommendation tasks); (2) a unified modular framework for developing and studying agentic recommender systems; and (3) the first comprehensive benchmark comparing 10 classical and agentic recommendation methods. Our findings demonstrate the superiority of agentic systems and establish actionable design guidelines for their core components. The benchmark environment has been rigorously validated through an open challenge and remains publicly available with a continuously maintained leaderboard~\footnote[2]{https://tsinghua-fib-lab.github.io/AgentSocietyChallenge/pages/overview.html}, fostering ongoing community engagement and reproducible research. The benchmark is available at: \hyperlink{https://huggingface.co/datasets/SGJQovo/AgentRecBench}{https://huggingface.co/datasets/SGJQovo/AgentRecBench}.

cs.IR↗

arXiv2025

Distributed Quantum Neural Networks on Distributed Photonic Quantum Computing

We introduce a distributed quantum-classical framework that synergizes photonic quantum neural networks (QNNs) with matrix-product-state (MPS) mapping to achieve parameter-efficient training of classical neural networks. By leveraging universal linear-optical decompositions of $M$-mode interferometers and photon-counting measurement statistics, our architecture generates neural parameters through a hybrid quantum-classical workflow: photonic QNNs with $M(M+1)/2$ trainable parameters produce high-dimensional probability distributions that are mapped to classical network weights via an MPS model with bond dimension $χ$. Empirical validation on MNIST classification demonstrates that photonic QT achieves an accuracy of $95.50\% \pm 0.84\%$ using 3,292 parameters ($χ= 10$), compared to $96.89\% \pm 0.31\%$ for classical baselines with 6,690 parameters. Moreover, a ten-fold compression ratio is achieved at $χ= 4$, with a relative accuracy loss of less than $3\%$. The framework outperforms classical compression techniques (weight sharing/pruning) by 6--12\% absolute accuracy while eliminating quantum hardware requirements during inference through classical deployment of compressed parameters. Simulations incorporating realistic photonic noise demonstrate the framework's robustness to near-term hardware imperfections. Ablation studies confirm quantum necessity: replacing photonic QNNs with random inputs collapses accuracy to chance level ($10.0\% \pm 0.5\%$). Photonic quantum computing's room-temperature operation, inherent scalability through spatial-mode multiplexing, and HPC-integrated architecture establish a practical pathway for distributed quantum machine learning, combining the expressivity of photonic Hilbert spaces with the deployability of classical neural networks.

quant-ph↗

arXiv2025

Nitrogen-Vacancy Engineering for Controlled Phase Transitions in CrN(111) Epitaxial Films

The phase transition in CrN epitaxial films is substantially suppressed by epitaxial constraint. Here, we propose that nitrogen (N) vacancies can be taken as a knob to regulate the phase transition of CrN(111) epitaxial films. To validate this concept, a series of CrN(111) films with controlled N concentrations (approximately from 0.0 to 5.0 at.%) were epitaxially grown on Al2O3(0001) substrates. Experimental characterization reveals that higher N vacancy concentrations significantly facilitate the out-of-plane contraction of the films at 273 K (0.8%), reaching up to 60% of the contraction magnitude of CrN powders (1.2%) without compromising the stability and reproducibility of the phase transition. Reducing N vacancy concentrations diminishes the lattice contraction, lowers the phase transition temperature to 193 K, and triggers a metallic to insulator transition in electrical behavior. First-principles calculations corroborate these findings, showing that N vacancies decrease the internal tensile stress within triangular Cr atomic layers, which enhances the out-of-plane contraction, elevates phase transition temperatures, and promotes bandgap closure. These results establish N vacancies as a critical factor governing phase transition dynamics in CrN systems and provide a practical strategy for successively engineering thermally responsive phase transitions in CrN films, advancing their potential for functional device applications.

cond-mat.mtrl-sci↗

arXiv2025

AgentSquare: Automatic LLM Agent Search in Modular Design Space

Recent advancements in Large Language Models (LLMs) have led to a rapid growth of agentic systems capable of handling a wide range of complex tasks. However, current research largely relies on manual, task-specific design, limiting their adaptability to novel tasks. In this paper, we introduce a new research problem: Modularized LLM Agent Search (MoLAS). We propose a modular design space that abstracts existing LLM agent designs into four fundamental modules with uniform IO interface: Planning, Reasoning, Tool Use, and Memory. Building on this design space, we present a novel LLM agent search framework called AgentSquare, which introduces two core mechanisms, i.e., module evolution and recombination, to efficiently search for optimized LLM agents. To further accelerate the process, we design a performance predictor that uses in-context surrogate models to skip unpromising agent designs. Extensive experiments across six benchmarks, covering the diverse scenarios of web, embodied, tool use and game applications, show that AgentSquare substantially outperforms hand-crafted agents, achieving an average performance gain of 17.2% against best-known human designs. Moreover, AgentSquare can generate interpretable design insights, enabling a deeper understanding of agentic architecture and its impact on task performance. We believe that the modular design space and AgentSquare search framework offer a platform for fully exploiting the potential of prior successful designs and consolidating the collective efforts of research community. Code repo is available at https://github.com/tsinghua-fib-lab/AgentSquare.

cs.CL↗

arXiv2025

AgentSociety Challenge: Designing LLM Agents for User Modeling and Recommendation on Web Platforms

The AgentSociety Challenge is the first competition in the Web Conference that aims to explore the potential of Large Language Model (LLM) agents in modeling user behavior and enhancing recommender systems on web platforms. The Challenge consists of two tracks: the User Modeling Track and the Recommendation Track. Participants are tasked to utilize a combined dataset from Yelp, Amazon, and Goodreads, along with an interactive environment simulator, to develop innovative LLM agents. The Challenge has attracted 295 teams across the globe and received over 1,400 submissions in total over the course of 37 official competition days. The participants have achieved 21.9% and 20.3% performance improvement for Track 1 and Track 2 in the Development Phase, and 9.1% and 15.9% in the Final Phase, representing a significant accomplishment. This paper discusses the detailed designs of the Challenge, analyzes the outcomes, and highlights the most successful LLM agent designs. To support further research and development, we have open-sourced the benchmark environment at https://tsinghua-fib-lab.github.io/AgentSocietyChallenge.

cs.IR↗

arXiv2025

A Large-scale Dataset with Behavior, Attributes, and Content of Mobile Short-video Platform

Short-video platforms show an increasing impact on people's daily lives nowadays, with billions of active users spending plenty of time each day. The interactions between users and online platforms give rise to many scientific problems across computational social science and artificial intelligence. However, despite the rapid development of short-video platforms, currently there are serious shortcomings in existing relevant datasets on three aspects: inadequate user-video feedback, limited user attributes and lack of video content. To address these problems, we provide a large-scale dataset with rich user behavior, attributes and video content from a real mobile short-video platform. This dataset covers 10,000 voluntary users and 153,561 videos, and we conduct four-fold technical validations of the dataset. First, we verify the richness of the behavior and attribute data. Second, we confirm the representing ability of the content features. Third, we provide benchmarking results on recommendation algorithms with our dataset. Finally, we explore the filter bubble phenomenon on the platform using the dataset. We believe the dataset could support the broad research community, including but not limited to user modeling, social science, human behavior understanding, etc. The dataset and code is available at https://github.com/tsinghua-fib-lab/ShortVideo_dataset.

cs.MM↗

arXiv2024

UrbanWorld: An Urban World Model for 3D City Generation

Cities, as the essential environment of human life, encompass diverse physical elements such as buildings, roads and vegetation, which continuously interact with dynamic entities like people and vehicles. Crafting realistic, interactive 3D urban environments is essential for nurturing AGI systems and constructing AI agents capable of perceiving, decision-making, and acting like humans in real-world environments. However, creating high-fidelity 3D urban environments usually entails extensive manual labor from designers, involving intricate detailing and representation of complex urban elements. Therefore, accomplishing this automatically remains a longstanding challenge. Toward this problem, we propose UrbanWorld, the first generative urban world model that can automatically create a customized, realistic and interactive 3D urban world with flexible control conditions. UrbanWorld incorporates four key stages in the generation pipeline: flexible 3D layout generation from OSM data or urban layout with semantic and height maps, urban scene design with Urban MLLM, controllable urban asset rendering via progressive 3D diffusion, and MLLM-assisted scene refinement. We conduct extensive quantitative analysis on five visual metrics, demonstrating that UrbanWorld achieves SOTA generation realism. Next, we provide qualitative results about the controllable generation capabilities of UrbanWorld using both textual and image-based prompts. Lastly, we verify the interactive nature of these environments by showcasing the agent perception and navigation within the created environments. We contribute UrbanWorld as an open-source tool available at https://github.com/Urban-World/UrbanWorld.

cs.CV↗

arXiv2024

Synergy-of-Thoughts: Eliciting Efficient Reasoning in Hybrid Language Models

Large language models (LLMs) have shown impressive emergent abilities in a wide range of tasks, but the associated expensive API cost greatly limits the real application. Previous works like chain-of-thought (CoT) and tree-of-thoughts (ToT) have predominately focused on enhancing accuracy, but overlook the rapidly increasing API cost, which could be particularly problematic for open-ended real-world tasks with huge solution spaces. Motivated by the dual process theory of human cognition, we propose "Synergy of Thoughts"(SoT) to unleash the synergistic potential of hybrid LLMs with different scales for efficient reasoning. By default, SoT uses smaller-scale language models to generate multiple low-cost intuitive thoughts, which resembles the parallel intuitions produced by System 1. We then design a confidence evaluator where the intuitive thoughts are cross-evaluated and introduce a controllable threshold mechanism to decide their mutual conflict. If these intuitive thoughts exhibit conflicts, SoT will invoke the reflective reasoning of scaled-up language models to emulate the intervention of System 2, which will override the intuitive thoughts and rectify the reasoning results. This framework is model-agnostic and training-free, which can be flexibly implemented with various off-the-shelf LLMs. Experiments on six representative reasoning tasks show that SoT substantially reduces the API cost by 38.3%-75.1%, and simultaneously achieves state-of-the-art reasoning accuracy and solution diversity. Notably, the average token cost reduction on open-ended tasks reaches up to 69.1%.

cs.CL↗

arXiv2024

Towards Biologically Plausible Computing: A Comprehensive Comparison

Backpropagation is a cornerstone algorithm in training neural networks for supervised learning, which uses a gradient descent method to update network weights by minimizing the discrepancy between actual and desired outputs. Despite its pivotal role in propelling deep learning advancements, the biological plausibility of backpropagation is questioned due to its requirements for weight symmetry, global error computation, and dual-phase training. To address this long-standing challenge, many studies have endeavored to devise biologically plausible training algorithms. However, a fully biologically plausible algorithm for training multilayer neural networks remains elusive, and interpretations of biological plausibility vary among researchers. In this study, we establish criteria for biological plausibility that a desirable learning algorithm should meet. Using these criteria, we evaluate a range of existing algorithms considered to be biologically plausible, including Hebbian learning, spike-timing-dependent plasticity, feedback alignment, target propagation, predictive coding, forward-forward algorithm, perturbation learning, local losses, and energy-based learning. Additionally, we empirically evaluate these algorithms across diverse network architectures and datasets. We compare the feature representations learned by these algorithms with brain activity recorded by non-invasive devices under identical stimuli, aiming to identify which algorithm can most accurately replicate brain activity patterns. We are hopeful that this study could inspire the development of new biologically plausible algorithms for training multilayer networks, thereby fostering progress in both the fields of neuroscience and machine learning.

cs.NE↗

arXiv2024

RNG: Reducing Multi-level Noise and Multi-grained Semantic Gap for Joint Multimodal Aspect-Sentiment Analysis

As an important multimodal sentiment analysis task, Joint Multimodal Aspect-Sentiment Analysis (JMASA), aiming to jointly extract aspect terms and their associated sentiment polarities from the given text-image pairs, has gained increasing concerns. Existing works encounter two limitations: (1) multi-level modality noise, i.e., instance- and feature-level noise; and (2) multi-grained semantic gap, i.e., coarse- and fine-grained gap. Both issues may interfere with accurate identification of aspect-sentiment pairs. To address these limitations, we propose a novel framework named RNG for JMASA. Specifically, to simultaneously reduce multi-level modality noise and multi-grained semantic gap, we design three constraints: (1) Global Relevance Constraint (GR-Con) based on text-image similarity for instance-level noise reduction, (2) Information Bottleneck Constraint (IB-Con) based on the Information Bottleneck (IB) principle for feature-level noise reduction, and (3) Semantic Consistency Constraint (SC-Con) based on mutual information maximization in a contrastive learning way for multi-grained semantic gap reduction. Extensive experiments on two datasets validate our new state-of-the-art performance.

cs.CL↗

arXiv2021

Genetic Meta-Structure Search for Recommendation on Heterogeneous Information Network

In the past decade, the heterogeneous information network (HIN) has become an important methodology for modern recommender systems. To fully leverage its power, manually designed network templates, i.e., meta-structures, are introduced to filter out semantic-aware information. The hand-crafted meta-structure rely on intense expert knowledge, which is both laborious and data-dependent. On the other hand, the number of meta-structures grows exponentially with its size and the number of node types, which prohibits brute-force search. To address these challenges, we propose Genetic Meta-Structure Search (GEMS) to automatically optimize meta-structure designs for recommendation on HINs. Specifically, GEMS adopts a parallel genetic algorithm to search meaningful meta-structures for recommendation, and designs dedicated rules and a meta-structure predictor to efficiently explore the search space. Finally, we propose an attention based multi-view graph convolutional network module to dynamically fuse information from different meta-structures. Extensive experiments on three real-world datasets suggest the effectiveness of GEMS, which consistently outperforms all baseline methods in HIN recommendation. Compared with simplified GEMS which utilizes hand-crafted meta-paths, GEMS achieves over $6\%$ performance gain on most evaluation metrics. More importantly, we conduct an in-depth analysis on the identified meta-structures, which sheds light on the HIN based recommender system design.

cs.IR↗

arXiv2012

EMRI data analysis with a phenomenological waveform

Extreme mass ratio inspirals (EMRIs) (capture and inspiral of a compact stellar mass object into a Massive Black Hole (MBH)) are among the most interesting objects for the gravitational wave astronomy. It is a very challenging task to detect those sources with the accurate estimation parameters of binaries primarily due to a large number of the secondary maxima on the likelihood surface. Search algorithms based on the matched filtering require computation of the gravitational waveform hundreds of thousands of times, which is currently not feasible with the most accurate (faithful) models of EMRIs. Here we propose to use a phenomenological template family which covers a large range of EMRIs parameter space. We use these phenomenological templates to detect the signal in the simulated data and then, assuming a particular EMRI model, estimate the physical parameters of the binary. We have separated the detection problem, which is done in a model-independent way, from the parameter estimation. For the latter one, we need to adopt the model for inspiral in order to map phenomenological parameters onto the physical parameter characterizing EMRIs.

gr-qc↗

arXiv2010

The search for spinning black hole binaries in mock LISA data using a genetic algorithm

Coalescing massive Black Hole binaries are the strongest and probably the most important gravitational wave sources in the LISA band. The spin and orbital precessions bring complexity in the waveform and make the likelihood surface richer in structure as compared to the non-spinning case. We introduce an extended multimodal genetic algorithm which utilizes the properties of the signal and the detector response function to analyze the data from the third round of mock LISA data challenge (MLDC 3.2). The performance of this method is comparable, if not better, to already existing algorithms. We have found all five sources present in MLDC 3.2 and recovered the coalescence time, chirp mass, mass ratio and sky location with reasonable accuracy. As for the orbital angular momentum and two spins of the Black Holes, we have found a large number of widely separated modes in the parameter space with similar maximum likelihood values.

gr-qc↗

arXiv2010

The Mock LISA Data Challenges: from Challenge 3 to Challenge 4

The Mock LISA Data Challenges are a program to demonstrate LISA data-analysis capabilities and to encourage their development. Each round of challenges consists of one or more datasets containing simulated instrument noise and gravitational waves from sources of undisclosed parameters. Participants analyze the datasets and report best-fit solutions for the source parameters. Here we present the results of the third challenge, issued in Apr 2008, which demonstrated the positive recovery of signals from chirping Galactic binaries, from spinning supermassive--black-hole binaries (with optimal SNRs between ~ 10 and 2000), from simultaneous extreme-mass-ratio inspirals (SNRs of 10-50), from cosmic-string-cusp bursts (SNRs of 10-100), and from a relatively loud isotropic background with Omega_gw(f) ~ 10^-11, slightly below the LISA instrument noise.

gr-qc↗

arXiv2009

The search for black hole binaries using a genetic algorithm

In this work we use genetic algorithm to search for the gravitational wave signal from the inspiralling massive Black Hole binaries in the simulated LISA data. We consider a single signal in the Gaussian instrumental noise. This is a first step in preparation for analysis of the third round of the mock LISA data challenge. We have extended a genetic algorithm utilizing the properties of the signal and the detector response function. The performance of this method is comparable, if not better, to already existing algorithms.

gr-qc↗

arXiv2009

The Mock LISA Data Challenges: from Challenge 1B to Challenge 3

The Mock LISA Data Challenges are a programme to demonstrate and encourage the development of LISA data-analysis capabilities, tools and techniques. At the time of this workshop, three rounds of challenges had been completed, and the next was about to start. In this article we provide a critical analysis of entries to the latest completed round, Challenge 1B. The entries confirm the consolidation of a range of data-analysis techniques for Galactic and massive--black-hole binaries, and they include the first convincing examples of detection and parameter estimation of extreme--mass-ratio inspiral sources. In this article we also introduce the next round, Challenge 3. Its data sets feature more realistic waveform models (e.g., Galactic binaries may now chirp, and massive--black-hole binaries may precess due to spin interactions), as well as new source classes (bursts from cosmic strings, isotropic stochastic backgrounds) and more complicated nonsymmetric instrument noise.

gr-qc↗

arXiv2007

Light Cone Structure near Null Infinity of the Kerr Metric

Motivated by our attempt to understand the question of angular momentum of a relativistic rotating source carried away by gravitational waves, in the asymptotic regime near future null infinity of the Kerr metric, a family of null hypersurfaces intersecting null infinity in shearfree (good) cuts are constructed by means of asymptotic expansion of the eikonal equation. The geometry of the null hypersurfaces as well asthe asymptotic structure of the Kerr metric near null infinity are studied. To the lowest order in angular momentum, the Bondi-Sachs form of the Kerr metric is worked out. The Newman-Unti formalism is then further developed, with which the Newman-Penrose constants of the Kerr metric are computed and shown to be zero. Possible physical implications of the vanishing of the Newman-Penrose constants of the Kerr metric are also briefly discussed.

gr-qc↗

arXiv2006

On Newman-Penrose constants of stationary space-times

We consider the general asymptotic expression of stationary space-time. Using Killing equation, we reduce the dynamical freedom of Einstein equation to the in-going gravitational wave $Ψ_0$. The general form of this function can be got. With the help of asymptotically algebraic special condition, we prove that all Newman-Penrose constants vanish.

gr-qc↗