arXiv subjects

An Luo

Publications and source records attributed to An Luo.

12 recordsLinked to original sources

arXiv2026

Nonuniformity Principle in Human-AI Coworking

As generative AI is increasingly applied to automate multi-step and high-stake workflows, human judgment and involvement remain essential for ensuring the quality of AI-generated outputs. In practice, while it is desirable for human experts to provide oversight on AI regularly, often by reviewing intermediate outputs, giving feedback, making corrections, and steering subsequent steps, such oversight is constrained by the time and resources that humans can afford. This creates a tension between the need for human oversight and AI's efficiency in delivering more output with less intervention. An important but underexplored question, then, is how to optimally engage humans in human-AI coworking. This work was originally motivated by our empirical observation that in long AI workflows, human oversight often improves user satisfaction while reducing unnecessary rework and token consumption. From there, we formulate the problem of where to place oversight stages in human-AI coworking. Under reasonable assumptions, we then develop the nonuniformity principle, which states that the optimal schedule places oversight stages with non-decreasing gaps along the workflow. We empirically validate this principle in two common AI agent workflows: writing literature reviews and constructing websites.

cs.AI↗

arXiv2026

Frequency Bias and OOD Generalization in Neural Operators under a Variable-Coefficient Wave Equation

Neural operators learn to map initial conditions to the terminal solution of partial differential equations (PDEs), providing a surrogate for the full operator mapping. This enables rapid prediction across different input configurations. While recent neural operator architectures have demonstrated strong performance on diverse PDE tasks, their behavior under structured distribution shifts remains insufficiently understood. To investigate this, we study operator learning in a wave propagation setting governed by a one-dimensional variable-coefficient wave equation, using two representative architectures, the Fourier Neural Operator (FNO) and the Deep Operator Network (DeepONet). To examine their generalization under distribution shifts, we consider structured out-of-distribution (OOD) settings that independently vary input frequency and coefficient smoothness. The results show that under smoothness shifts, both models maintain stable performance, with FNO achieving lower error. In contrast, under frequency shifts, FNO exhibits a sharp increase in error under unseen high-frequency inputs, whereas DeepONet shows milder degradation despite higher overall error. Our analysis reveals that these differences arise from how each architecture represents and responds to variations in frequency structure. Together, these findings highlight a fundamental gap between strong in-distribution performance and generalization under distribution shifts in operator learning, underscoring the role of architectural representation bias in developing more reliable neural operators for physics-based PDE simulations beyond the training distribution.

cs.LG↗

arXiv2026

ADD for Multi-Bit Image Watermarking

As generative models enable rapid creation of high-fidelity images, societal concerns about misinformation and authenticity have intensified. A promising remedy is multi-bit image watermarking, which embeds a multi-bit message into an image so that a verifier can later detect whether the image is generated by someone and further identify the source by decoding the embedded message. Existing approaches often fall short in capacity, resilience to common image distortions, and theoretical justification. To address these limitations, we propose ADD (Add, Dot, Decode), a multi-bit image watermarking method with two stages: learning a watermark to be linearly combined with the multi-bit message and added to the image, and decoding through inner products between the watermarked image and the learned watermark. On the standard MS-COCO benchmark, we demonstrate that for the challenging task of 48-bit watermarking, ADD achieves 100\% decoding accuracy, with performance dropping by at most 2\% under a wide range of image distortions, substantially smaller than the 14\% average drop of state-of-the-art methods. In addition, ADD achieves substantial computational gains, with 2-fold faster embedding and 7.4-fold faster decoding than the fastest existing method. We further provide a theoretical analysis explaining why the learned watermark and the corresponding decoding rule are effective.

stat.ML↗

arXiv2026

AgentDS Technical Report: Benchmarking the Future of Human-AI Collaboration in Domain-Specific Data Science

Data science plays a critical role in transforming complex data into actionable insights across numerous domains. Recent developments in large language models (LLMs) and artificial intelligence (AI) agents have significantly automated data science workflow. However, it remains unclear to what extent AI agents can match the performance of human experts on domain-specific data science tasks, and in which aspects human expertise continues to provide advantages. We introduce AgentDS, a benchmark and competition designed to evaluate both AI agents and human-AI collaboration performance in domain-specific data science. AgentDS consists of 17 challenges across six industries: commerce, food production, healthcare, insurance, manufacturing, and retail banking. We conducted an open competition involving 29 teams and 80 participants, enabling systematic comparison between human-AI collaborative approaches and AI-only baselines. Our results show that current AI agents struggle with domain-specific reasoning. AI-only baselines perform below the top quartile of competition participants, while the strongest solutions arise from human-AI collaboration. These findings challenge the narrative of complete automation by AI and underscore the enduring importance of human expertise in data science, while illuminating directions for the next generation of AI. Visit the AgentDS website here: https://agentds.org/ and open source datasets here: https://huggingface.co/datasets/lainmn/AgentDS .

cs.LG↗

arXiv2025

Can Agentic AI Match the Performance of Human Data Scientists?

Data science plays a critical role in transforming complex data into actionable insights across numerous domains. Recent developments in large language models (LLMs) have significantly automated data science workflows, but a fundamental question persists: Can these agentic AI systems truly match the performance of human data scientists who routinely leverage domain-specific knowledge? We explore this question by designing a prediction task where a crucial latent variable is hidden in relevant image data instead of tabular features. As a result, agentic AI that generates generic codes for modeling tabular data cannot perform well, while human experts could identify the important hidden variable using domain knowledge. We demonstrate this idea with a synthetic dataset for property insurance. Our experiments show that agentic AI that relies on generic analytics workflow falls short of methods that use domain-specific insights. This highlights a key limitation of the current agentic AI for data science and underscores the need for future research to develop agentic AI systems that can better recognize and incorporate domain knowledge.

cs.LG↗

arXiv2025

AssistedDS: Benchmarking How External Domain Knowledge Assists LLMs in Automated Data Science

Large language models (LLMs) have advanced the automation of data science workflows. Yet it remains unclear whether they can critically leverage external domain knowledge as human data scientists do in practice. To answer this question, we introduce AssistedDS (Assisted Data Science), a benchmark designed to systematically evaluate how LLMs handle domain knowledge in tabular prediction tasks. AssistedDS features both synthetic datasets with explicitly known generative mechanisms and real-world Kaggle competitions, each accompanied by curated bundles of helpful and adversarial documents. These documents provide domain-specific insights into data cleaning, feature engineering, and model selection. We assess state-of-the-art LLMs on their ability to discern and apply beneficial versus harmful domain knowledge, evaluating submission validity, information recall, and predictive performance. Our results demonstrate three key findings: (1) LLMs frequently exhibit an uncritical adoption of provided information, significantly impairing their predictive performance when adversarial content is introduced, (2) helpful guidance is often insufficient to counteract the negative influence of adversarial information, and (3) in Kaggle datasets, LLMs often make errors in handling time-series data, applying consistent feature engineering across different folds, and interpreting categorical variables correctly. These findings highlight a substantial gap in current models' ability to critically evaluate and leverage expert knowledge, underscoring an essential research direction for developing more robust, knowledge-aware automated data science systems. Our data and code are publicly available here: https://github.com/jeremyxianx/Assisted-DS

cs.LG↗

arXiv2025

An Outlook on the Opportunities and Challenges of Multi-Agent AI Systems

A multi-agent AI system (MAS) is composed of multiple autonomous agents that interact, exchange information, and make decisions based on internal generative models. Recent advances in large language models and tool-using agents have made MAS increasingly practical in areas like scientific discovery and collaborative automation. However, key questions remain: When are MAS more effective than single-agent systems? What new safety risks arise from agent interactions? And how should we evaluate their reliability and structure? This paper outlines a formal framework for analyzing MAS, focusing on two core aspects: effectiveness and safety. We explore whether MAS truly improve robustness, adaptability, and performance, or merely repackage known techniques like ensemble learning. We also study how inter-agent dynamics may amplify or suppress system vulnerabilities. While MAS are relatively new to the signal processing community, we envision them as a powerful abstraction that extends classical tools like distributed estimation and sensor fusion to higher-level, policy-driven inference. Through experiments on data science automation, we highlight the potential of MAS to reshape how signal processing systems are designed and trusted.

cs.MA↗

arXiv2025

Ice Cream Doesn't Cause Drowning: Benchmarking LLMs Against Statistical Pitfalls in Causal Inference

Reliable causal inference is essential for making decisions in high-stakes areas like medicine, economics, and public policy. However, it remains unclear whether large language models (LLMs) can handle rigorous and trustworthy statistical causal inference. Current benchmarks usually involve simplified tasks. For example, these tasks might only ask LLMs to identify semantic causal relationships or draw conclusions directly from raw data. As a result, models may overlook important statistical pitfalls, such as Simpson's paradox or selection bias. This oversight limits the applicability of LLMs in the real world. To address these limitations, we propose CausalPitfalls, a comprehensive benchmark designed to rigorously evaluate the capability of LLMs in overcoming common causal inference pitfalls. Our benchmark features structured challenges across multiple difficulty levels, each paired with grading rubrics. This approach allows us to quantitatively measure both causal reasoning capabilities and the reliability of LLMs' responses. We evaluate models using two protocols: (1) direct prompting, which assesses intrinsic causal reasoning, and (2) code-assisted prompting, where models generate executable code for explicit statistical analysis. Additionally, we validate the effectiveness of this judge by comparing its scoring with assessments from human experts. Our results reveal significant limitations in current LLMs when performing statistical causal inference. The CausalPitfalls benchmark provides essential guidance and quantitative metrics to advance the development of trustworthy causal reasoning systems.

cs.AI↗

arXiv2024

Microscopic Study on Superexchange Dynamics of Composite Spin-1 Bosons

We report on an experimental simulation of the spin-1 Heisenberg model with composite bosons in a one-dimensional chain based on the two-component Bose-Hubbard model. Exploiting our site-and spin-resolved quantum gas microscope, we observed faster superexchange dynamics of the spin-1 system compared to its spin-1/2 counterpart, which is attributed to the enhancement effect of multi-bosons. We further probed the non-equilibrium spin dynamics driven by the superexchange and single-ion anisotropy terms, unveiling the linear expansion of the spin-spin correlations, which is limited by the Lieb-Robinson bound. Based on the superexchange process, we prepared and verified the entangled qutrits pairs with these composite spin-1 bosons, potentially being applied in qutrit-based quantum information processing.

cond-mat.quant-gas↗

arXiv2024

Observation of counterflow superfluidity in a two-component Mott insulator

The counterflow superfluidity (CSF) was predicted two decades ago. Counterintuitively, while both components in the CSF have fluidity, their correlated counterflow currents cancel out leading the overall system to an incompressible Mott insulator. However, realizing and identifying the CSF remain challenging due to the request on extreme experimental capabilities in a single setup. Here, we observe the CSF in a binary Bose mixture in optical lattices. We prepare a low-entropy spin-Mott state by conveying and merging two spin-1/2 bosonic atoms at every site and drive it adiabatically to the CSF at $\sim$ 1 nK. Antipair correlations of the CSF are probed though a site- and spin-resolved quantum gas microscope in both real and momentum spaces. These techniques and observations provide accessibility to the symmetry-protected topological quantum matters.

cond-mat.quant-gas↗

arXiv2022

Quantifying Quantum Computational Advantage on a Processor of Ultracold Atoms

Nonequilibrium dynamics of quantum many-body systems is challenging for classical computing, providing opportunities for demonstrating practical quantum computational advantage with analogue quantum simulators. Owing to the intimate connection with a random matrix ensemble, it is proposed to be classically intractable to sample the driven thermalized many-body states of a Bose-Hubbard system, and further extract multi-point correlations from the output-strings for characterizing quantum systems. Here, leveraging dedicated precise manipulations and atom-number-resolved detection through a quantum gas microscope with bichromatic superlattices, we perform sampling of the driven Hubbard chains and two-leg ladders in the thermalized phase involving up to 64 sites with 20 atoms, yielding a Hilbert space dimension of $10^{19}$ and outpacing the most powerful supercomputer in terms of sampling rate by three orders of magnitude. The volume law scaling of the \Renyi entanglement entropy in the thermalized phase is observed, which hinders efficient classical simulation for large systems. We employ the Bayesian tests to verify that our prepared systems operate in the driven thermalized phase. Multi-point correlations of up to 14th-order extracted from the experimental samples offer clear distinctions between the thermalized and many-body-localized phases, where classical computations such as tensor network fails to give accurate and faithful predictions within a reasonable time cost. Our work demonstrates the sampling of a interacting chaotic system performed on a quantum processor of ultracold atoms and opens the door of utilizable quantum computational advantage in simulating Floquet dynamics of many-body systems.

cond-mat.quant-gas↗

arXiv2022

Functional building blocks for scalable multipartite entanglement in optical lattices

Featuring excellent coherence and operated parallelly, ultracold atoms in optical lattices form a competitive candidate for quantum computation. For this, a massive number of parallel entangled atom pairs have been realized in superlattices. However, the more formidable challenge is to scale-up and detect multipartite entanglement due to the lack of manipulations over local atomic spins in retro-reflected bichromatic superlattices. Here we developed a new architecture based on a cross-angle spin-dependent superlattice for implementing layers of quantum gates over moderately-separated atoms incorporated with a quantum gas microscope for single-atom manipulation. We created and verified functional building blocks for scalable multipartite entanglement by connecting Bell pairs to one-dimensional 10-atom chains and two-dimensional plaquettes of $2\times4$ atoms. This offers a new platform towards scalable quantum computation and simulation.

cond-mat.quant-gas↗