arXiv subjects

Tian Lan

Publications and source records attributed to Tian Lan.

At least 145 records · Page 8Linked to original sources

arXiv2024

RGMDT: Return-Gap-Minimizing Decision Tree Extraction in Non-Euclidean Metric Space

Deep Reinforcement Learning (DRL) algorithms have achieved great success in solving many challenging tasks while their black-box nature hinders interpretability and real-world applicability, making it difficult for human experts to interpret and understand DRL policies. Existing works on interpretable reinforcement learning have shown promise in extracting decision tree (DT) based policies from DRL policies with most focus on the single-agent settings while prior attempts to introduce DT policies in multi-agent scenarios mainly focus on heuristic designs which do not provide any quantitative guarantees on the expected return. In this paper, we establish an upper bound on the return gap between the oracle expert policy and an optimal decision tree policy. This enables us to recast the DT extraction problem into a novel non-euclidean clustering problem over the local observation and action values space of each agent, with action values as cluster labels and the upper bound on the return gap as clustering loss. Both the algorithm and the upper bound are extended to multi-agent decentralized DT extractions by an iteratively-grow-DT procedure guided by an action-value function conditioned on the current DTs of other agents. Further, we propose the Return-Gap-Minimization Decision Tree (RGMDT) algorithm, which is a surprisingly simple design and is integrated with reinforcement learning through the utilization of a novel Regularized Information Maximization loss. Evaluations on tasks like D4RL show that RGMDT significantly outperforms heuristic DT-based baselines and can achieve nearly optimal returns under given DT complexity constraints (e.g., maximum number of DT nodes).

cs.LG↗

arXiv2024

CriticEval: Evaluating Large Language Model as Critic

Critique ability, i.e., the capability of Large Language Models (LLMs) to identify and rectify flaws in responses, is crucial for their applications in self-improvement and scalable oversight. While numerous studies have been proposed to evaluate critique ability of LLMs, their comprehensiveness and reliability are still limited. To overcome this problem, we introduce CriticEval, a novel benchmark designed to comprehensively and reliably evaluate critique ability of LLMs. Specifically, to ensure the comprehensiveness, CriticEval evaluates critique ability from four dimensions across nine diverse task scenarios. It evaluates both scalar-valued and textual critiques, targeting responses of varying quality. To ensure the reliability, a large number of critiques are annotated to serve as references, enabling GPT-4 to evaluate textual critiques reliably. Extensive evaluations of open-source and closed-source LLMs first validate the reliability of evaluation in CriticEval. Then, experimental results demonstrate the promising potential of open-source LLMs, the effectiveness of critique datasets and several intriguing relationships between the critique ability and some critical factors, including task types, response qualities and critique dimensions.

cs.CL↗

arXiv2024

Training Language Models to Critique With Multi-agent Feedback

Critique ability, a meta-cognitive capability of humans, presents significant challenges for LLMs to improve. Recent works primarily rely on supervised fine-tuning (SFT) using critiques generated by a single LLM like GPT-4. However, these model-generated critiques often exhibit flaws due to the inherent complexity of the critique. Consequently, fine-tuning LLMs on such flawed critiques typically limits the model's performance and propagates these flaws into the learned model. To overcome these challenges, this paper proposes a novel data generation pipeline, named MultiCritique, that improves the critique ability of LLMs by utilizing multi-agent feedback in both the SFT and reinforcement learning (RL) stages. First, our data generation pipeline aggregates high-quality critiques from multiple agents instead of a single model, with crucial information as input for simplifying the critique. Furthermore, our pipeline improves the preference accuracy of critique quality through multi-agent feedback, facilitating the effectiveness of RL in improving the critique ability of LLMs. Based on our proposed MultiCritique data generation pipeline, we construct the MultiCritiqueDataset for the SFT and RL fine-tuning stages. Extensive experimental results on two benchmarks demonstrate: 1) the superior quality of our constructed SFT dataset compared to existing critique datasets; 2) additional improvements to the critique ability of LLMs brought by the RL stage. Notably, our fine-tuned 7B model significantly surpasses other advanced 7B-13B open-source models, approaching the performance of advanced 70B LLMs and GPT-4. Codes, datasets and model weights will be publicly available.

cs.CL↗

arXiv2024

Learning-based Two-tiered Online Optimization of Region-wide Datacenter Resource Allocation

Online optimization of resource management for large-scale data centers and infrastructures to meet dynamic capacity reservation demands and various practical constraints (e.g., feasibility and robustness) is a very challenging problem. Mixed Integer Programming (MIP) approaches suffer from recognized limitations in such a dynamic environment, while learning-based approaches may face with prohibitively large state/action spaces. To this end, this paper presents a novel two-tiered online optimization to enable a learning-based Resource Allowance System (RAS). To solve optimal server-to-reservation assignment in RAS in an online fashion, the proposed solution leverages a reinforcement learning (RL) agent to make high-level decisions, e.g., how much resource to select from the Main Switch Boards (MSBs), and then a low-level Mixed Integer Linear Programming (MILP) solver to generate the local server-to-reservation mapping, conditioned on the RL decisions. We take into account fault tolerance, server movement minimization, and network affinity requirements and apply the proposed solution to large-scale RAS problems. To provide interpretability, we further train a decision tree model to explain the learned policies and to prune unreasonable corner cases at the low-level MILP solver, resulting in further performance improvement. Extensive evaluations show that our two-tiered solution outperforms baselines such as pure MIP solver by over $15\%$ while delivering $100\times$ speedup in computation.

cs.NI↗

arXiv2024

The cloud cover and meteorological parameters at the Lenghu site on the Tibetan Plateau

The cloud cover and meteorological parameters serve as fundamental criteria for the qualification of an astronomical observatory working in optical and infrared wavelengths. In this paper, we present a systematic assessment of key meteorological parameters at the Lenghu site. The datasets adopted in this study includes the meteorological parameters collected at the local weather stations at the site and in the Lenghu Town, the sky brightness at the local zenith acquired by the Sky Quality Meters and night sky all-sky images from a digital camera, the ERA5 reanalysis database and global climate monitoring data. From 2019 to 2023, the fractional observable time of photometric condition is 69.70%, 74.97%, 70.26%, 74.27% and 65.12%, respectively. The fractional observing time is inversely correlated with surface air temperature, relative humidity, precipitable water vapor, and dew temperature, demonstrating that the observing conditions are influenced by these meteorological parameters. Large-scale air-sea interactions affect the climate at Lenghu site, which in fact delivers a clue to understand the irregularity of 2023. Specifically, precipitable water vapor at Lenghu site is correlated to both the westerly wind index and the summer North Atlantic Oscillation index, the yearly average temperature of Lenghu site is observed to increase significantly during the occurrence of a strong El Niño event and the relative humidity anomaly at Lenghu site is correlated to the Pacific Decadal Oscillation index. The decrease of fractional observing time in 2023 was due to the ongoing strong El Niño event and relevant global climate change. We underscore the substantial role of global climate change in regulating astronomical observing conditions and the necessity for long-term continuous monitoring of the astronomical meteorological parameters at Lenghu site.

astro-ph.IM↗

arXiv2024

DiffuTraj: A Stochastic Vessel Trajectory Prediction Approach via Guided Diffusion Process

Maritime vessel maneuvers, characterized by their inherent complexity and indeterminacy, requires vessel trajectory prediction system capable of modeling the multi-modality nature of future motion states. Conventional stochastic trajectory prediction methods utilize latent variables to represent the multi-modality of vessel motion, however, tends to overlook the complexity and dynamics inherent in maritime behavior. In contrast, we explicitly simulate the transition of vessel motion from uncertainty towards a state of certainty, effectively handling future indeterminacy in dynamic scenes. In this paper, we present a novel framework (\textit{DiffuTraj}) to conceptualize the trajectory prediction task as a guided reverse process of motion pattern uncertainty diffusion, in which we progressively remove uncertainty from maritime regions to delineate the intended trajectory. Specifically, we encode the previous states of the target vessel, vessel-vessel interactions, and the environment context as guiding factors for trajectory generation. Subsequently, we devise a transformer-based conditional denoiser to capture spatio-temporal dependencies, enabling the generation of trajectories better aligned for particular maritime environment. Comprehensive experiments on vessel trajectory prediction benchmarks demonstrate the superiority of our method.

cs.CV↗

arXiv2024

xLAM: A Family of Large Action Models to Empower AI Agent Systems

Autonomous agents powered by large language models (LLMs) have attracted significant research interest. However, the open-source community faces many challenges in developing specialized models for agent tasks, driven by the scarcity of high-quality agent datasets and the absence of standard protocols in this area. We introduce and publicly release xLAM, a series of large action models designed for AI agent tasks. The xLAM series includes five models with both dense and mixture-of-expert architectures, ranging from 1B to 8x22B parameters, trained using a scalable, flexible pipeline that unifies, augments, and synthesizes diverse datasets to enhance AI agents' generalizability and performance across varied environments. Our experimental results demonstrate that xLAM consistently delivers exceptional performance across multiple agent ability benchmarks, notably securing the 1st position on the Berkeley Function-Calling Leaderboard, outperforming GPT-4, Claude-3, and many other models in terms of tool use. By releasing the xLAM series, we aim to advance the performance of open-source LLMs for autonomous AI agents, potentially accelerating progress and democratizing access to high-performance models for agent tasks. Models are available at https://huggingface.co/collections/Salesforce/xlam-models-65f00e2a0a63bbcd1c2dade4

cs.CL↗

arXiv2024

In-Memory Learning Automata Architecture using Y-Flash Cell

The modern implementation of machine learning architectures faces significant challenges due to frequent data transfer between memory and processing units. In-memory computing, primarily through memristor-based analog computing, offers a promising solution to overcome this von Neumann bottleneck. In this technology, data processing and storage are located inside the memory. Here, we introduce a novel approach that utilizes floating-gate Y-Flash memristive devices manufactured with a standard 180 nm CMOS process. These devices offer attractive features, including analog tunability and moderate device-to-device variation; such characteristics are essential for reliable decision-making in ML applications. This paper uses a new machine learning algorithm, the Tsetlin Machine (TM), for in-memory processing architecture. The TM's learning element, Automaton, is mapped into a single Y-Flash cell, where the Automaton's range is transferred into the Y-Flash's conductance scope. Through comprehensive simulations, the proposed hardware implementation of the learning automata, particularly for Tsetlin machines, has demonstrated enhanced scalability and on-edge learning capabilities.

cs.AR↗

arXiv2024

Diversity Empowers Intelligence: Integrating Expertise of Software Engineering Agents

Large language model (LLM) agents have shown great potential in solving real-world software engineering (SWE) problems. The most advanced open-source SWE agent can resolve over 27% of real GitHub issues in SWE-Bench Lite. However, these sophisticated agent frameworks exhibit varying strengths, excelling in certain tasks while underperforming in others. To fully harness the diversity of these agents, we propose DEI (Diversity Empowered Intelligence), a framework that leverages their unique expertise. DEI functions as a meta-module atop existing SWE agent frameworks, managing agent collectives for enhanced problem-solving. Experimental results show that a DEI-guided committee of agents is able to surpass the best individual agent's performance by a large margin. For instance, a group of open-source SWE agents, with a maximum individual resolve rate of 27.3% on SWE-Bench Lite, can achieve a 34.3% resolve rate with DEI, making a 25% improvement and beating most closed-source solutions. Our best-performing group excels with a 55% resolve rate, securing the highest ranking on SWE-Bench Lite. Our findings contribute to the growing body of research on collaborative AI systems and their potential to solve complex software engineering challenges.

cs.SE↗

arXiv2024

Enabling High Data Throughput Reinforcement Learning on GPUs: A Domain Agnostic Framework for Data-Driven Scientific Research

We introduce WarpSci, a domain agnostic framework designed to overcome crucial system bottlenecks encountered in the application of reinforcement learning to intricate environments with vast datasets featuring high-dimensional observation or action spaces. Notably, our framework eliminates the need for data transfer between the CPU and GPU, enabling the concurrent execution of thousands of simulations on a single or multiple GPUs. This high data throughput architecture proves particularly advantageous for data-driven scientific research, where intricate environment models are commonly essential.

cs.LG↗

arXiv2024

APIGen: Automated Pipeline for Generating Verifiable and Diverse Function-Calling Datasets

The advancement of function-calling agent models requires diverse, reliable, and high-quality datasets. This paper presents APIGen, an automated data generation pipeline designed to synthesize verifiable high-quality datasets for function-calling applications. We leverage APIGen and collect 3,673 executable APIs across 21 different categories to generate diverse function-calling datasets in a scalable and structured manner. Each data in our dataset is verified through three hierarchical stages: format checking, actual function executions, and semantic verification, ensuring its reliability and correctness. We demonstrate that models trained with our curated datasets, even with only 7B parameters, can achieve state-of-the-art performance on the Berkeley Function-Calling Benchmark, outperforming multiple GPT-4 models. Moreover, our 1B model achieves exceptional performance, surpassing GPT-3.5-Turbo and Claude-3 Haiku. We release a dataset containing 60,000 high-quality entries, aiming to advance the field of function-calling agent domains. The dataset is available on Huggingface: https://huggingface.co/datasets/Salesforce/xlam-function-calling-60k and the project homepage: https://apigen-pipeline.github.io/

cs.CL↗

arXiv2024

Learning from Random Demonstrations: Offline Reinforcement Learning with Importance-Sampled Diffusion Models

Generative models such as diffusion have been employed as world models in offline reinforcement learning to generate synthetic data for more effective learning. Existing work either generates diffusion models one-time prior to training or requires additional interaction data to update it. In this paper, we propose a novel approach for offline reinforcement learning with closed-loop policy evaluation and world-model adaptation. It iteratively leverages a guided diffusion world model to directly evaluate the offline target policy with actions drawn from it, and then performs an importance-sampled world model update to adaptively align the world model with the updated policy. We analyzed the performance of the proposed method and provided an upper bound on the return gap between our method and the real environment under an optimal policy. The result sheds light on various factors affecting learning performance. Evaluations in the D4RL environment show significant improvement over state-of-the-art baselines, especially when only random or medium-expertise demonstrations are available -- thus requiring improved alignment between the world model and offline policy evaluation.

cs.LG↗

arXiv2024

Deep Generative Models for Offline Policy Learning: Tutorial, Survey, and Perspectives on Future Directions

Deep generative models (DGMs) have demonstrated great success across various domains, particularly in generating texts, images, and videos using models trained from offline data. Similarly, data-driven decision-making and robotic control also necessitate learning a generator function from the offline data to serve as the strategy or policy. In this case, applying deep generative models in offline policy learning exhibits great potential, and numerous studies have explored in this direction. However, this field still lacks a comprehensive review and so developments of different branches are relatively independent. In this paper, we provide the first systematic review on the applications of deep generative models for offline policy learning. In particular, we cover five mainstream deep generative models, including Variational Auto-Encoders, Generative Adversarial Networks, Normalizing Flows, Transformers, and Diffusion Models, and their applications in both offline reinforcement learning (offline RL) and imitation learning (IL). Offline RL and IL are two main branches of offline policy learning and are widely-adopted techniques for sequential decision-making. Notably, for each type of DGM-based offline policy learning, we distill its fundamental scheme, categorize related works based on the usage of the DGM, and sort out the development process of algorithms in that field. Subsequent to the main content, we provide in-depth discussions on deep generative models and offline policy learning as a summary, based on which we present our perspectives on future research directions. This work offers a hands-on reference for the research progress in deep generative models for offline policy learning, and aims to inspire improved DGM-based offline RL or IL algorithms. For convenience, we maintain a paper list on https://github.com/LucasCJYSDL/DGMs-for-Offline-Policy-Learning.

cs.LG↗

arXiv2024

Modeling Other Players with Bayesian Beliefs for Games with Incomplete Information

Bayesian games model interactive decision-making where players have incomplete information -- e.g., regarding payoffs and private data on players' strategies and preferences -- and must actively reason and update their belief models (with regard to such information) using observation and interaction history. Existing work on counterfactual regret minimization have shown great success for games with complete or imperfect information, but not for Bayesian games. To this end, we introduced a new CFR algorithm: Bayesian-CFR and analyze its regret bound with respect to Bayesian Nash Equilibria in Bayesian games. First, we present a method for updating the posterior distribution of beliefs about the game and other players' types. The method uses a kernel-density estimate and is shown to converge to the true distribution. Second, we define Bayesian regret and present a Bayesian-CFR minimization algorithm for computing the Bayesian Nash equilibrium. Finally, we extend this new approach to other existing algorithms, such as Bayesian-CFR+ and Deep Bayesian CFR. Experimental results show that our proposed solutions significantly outperform existing methods in classical Texas Hold'em games.

cs.GT↗

arXiv2024

Monocular Gaussian SLAM with Language Extended Loop Closure

Recently,3DGaussianSplattinghasshowngreatpotentialin visual Simultaneous Localization And Mapping (SLAM). Existing methods have achieved encouraging results on RGB-D SLAM, but studies of the monocular case are still scarce. Moreover, they also fail to correct drift errors due to the lack of loop closure and global optimization. In this paper, we present MG-SLAM, a monocular Gaussian SLAM with a language-extended loop closure module capable of performing drift-corrected tracking and high-fidelity reconstruction while achieving a high-level understanding of the environment. Our key idea is to represent the global map as 3D Gaussian and use it to guide the estimation of the scene geometry, thus mitigating the efforts of missing depth information. Further, an additional language-extended loop closure module which is based on CLIP feature is designed to continually perform global optimization to correct drift errors accumulated as the system runs. Our system shows promising results on multiple challenging datasets in both tracking and mapping and even surpasses some existing RGB-D methods.

cs.CV↗

arXiv2024

On a class of fusion 2-category symmetry: condensation completion of braided fusion category

Recently, many studies are focused on generalized global symmetry, a mixture of both invertible and non-invertible symmetries in various space-time dimensions. The complete structure of generalized global symmetry is described by higher fusion category theory. In this paper, We first review the construction of fusion 2-category symmetry $Σ\cal B$ where $\cal B$ is a a braided fusion category. In particular, we elaborate on the monoidal structure of $Σ\cal B$ which determines fusion rules and controls the dynamics of topological operators/defects. We then take $Σ\mathrm{sVec}$ as an example to demonstrate how we calculate fusion rule, quantum dimension and 10j-symbol of the fusion 2-category. With our algorithm, all these data can be efficiently encoded and computed in computer program. The complete program will be uploaded to github soon. Our work can be thought as explicitly computing the representation theory of $\cal B$, in analogy to, for example the representation theory of $SU(2)$. The choice of basis bimodule maps are in analogy to the Clebsch-Gordon coefficients and the 10j-symbol are in analogy to the 6j-symbol.

hep-th↗

arXiv2024

SwiftRL: Towards Efficient Reinforcement Learning on Real Processing-In-Memory Systems

Reinforcement Learning (RL) trains agents to learn optimal behavior by maximizing reward signals from experience datasets. However, RL training often faces memory limitations, leading to execution latencies and prolonged training times. To overcome this, SwiftRL explores Processing-In-Memory (PIM) architectures to accelerate RL workloads. We achieve near-linear performance scaling by implementing RL algorithms like Tabular Q-learning and SARSA on UPMEM PIM systems and optimizing for hardware. Our experiments on OpenAI GYM environments using UPMEM hardware demonstrate superior performance compared to CPU and GPU implementations.

cs.LG↗

arXiv2024

MultiFun-DAG: Multivariate Functional Directed Acyclic Graph

Directed Acyclic Graphical (DAG) models efficiently formulate causal relationships in complex systems. Traditional DAGs assume nodes to be scalar variables, characterizing complex systems under a facile and oversimplified form. This paper considers that nodes can be multivariate functional data and thus proposes a multivariate functional DAG (MultiFun-DAG). It constructs a hidden bilinear multivariate function-to-function regression to describe the causal relationships between different nodes. Then an Expectation-Maximum algorithm is used to learn the graph structure as a score-based algorithm with acyclic constraints. Theoretical properties are diligently derived. Prudent numerical studies and a case study from urban traffic congestion analysis are conducted to show MultiFun-DAG's effectiveness.

stat.ME↗