Search arXiv⌕ Search

arXiv subjects

Yang Yu

Publications and source records attributed to Yang Yu.

At least 271 records · Page 15Linked to original sources
arXiv2023

WeLM: A Well-Read Pre-trained Language Model for Chinese

Large Language Models pre-trained with self-supervised learning have demonstrated impressive zero-shot generalization capabilities on a wide spectrum of tasks. In this work, we present WeLM: a well-read pre-trained language model for Chinese that is able to seamlessly perform different types of tasks with zero or few-shot demonstrations. WeLM is trained with 10B parameters by "reading" a curated high-quality corpus covering a wide range of topics. We show that WeLM is equipped with broad knowledge on various domains and languages. On 18 monolingual (Chinese) tasks, WeLM can significantly outperform existing pre-trained models with similar sizes and match the performance of models up to 25 times larger. WeLM also exhibits strong capabilities in multi-lingual and code-switching understanding, outperforming existing multilingual language models pre-trained on 30 languages. Furthermore, We collected human-written prompts for a large set of supervised datasets in Chinese and fine-tuned WeLM with multi-prompted training. The resulting model can attain strong generalization on unseen types of tasks and outperform the unsupervised WeLM in zero-shot learning. Finally, we demonstrate that WeLM has basic skills at explaining and calibrating the decisions from itself, which can be promising directions for future research. Our models can be applied from https://welm.weixin.qq.com/docs/api/.

cs.CL↗
arXiv2023

Robust multi-agent coordination via evolutionary generation of auxiliary adversarial attackers

Cooperative multi-agent reinforcement learning (CMARL) has shown to be promising for many real-world applications. Previous works mainly focus on improving coordination ability via solving MARL-specific challenges (e.g., non-stationarity, credit assignment, scalability), but ignore the policy perturbation issue when testing in a different environment. This issue hasn't been considered in problem formulation or efficient algorithm design. To address this issue, we firstly model the problem as a limited policy adversary Dec-POMDP (LPA-Dec-POMDP), where some coordinators from a team might accidentally and unpredictably encounter a limited number of malicious action attacks, but the regular coordinators still strive for the intended goal. Then, we propose Robust Multi-Agent Coordination via Evolutionary Generation of Auxiliary Adversarial Attackers (ROMANCE), which enables the trained policy to encounter diversified and strong auxiliary adversarial attacks during training, thus achieving high robustness under various policy perturbations. Concretely, to avoid the ego-system overfitting to a specific attacker, we maintain a set of attackers, which is optimized to guarantee the attackers high attacking quality and behavior diversity. The goal of quality is to minimize the ego-system coordination effect, and a novel diversity regularizer based on sparse action is applied to diversify the behaviors among attackers. The ego-system is then paired with a population of attackers selected from the maintained attacker set, and alternately trained against the constantly evolving attackers. Extensive experiments on multiple scenarios from SMAC indicate our ROMANCE provides comparable or better robustness and generalization ability than other baselines.

cs.MA↗
arXiv2023

Fast Teammate Adaptation in the Presence of Sudden Policy Change

In cooperative multi-agent reinforcement learning (MARL), where an agent coordinates with teammate(s) for a shared goal, it may sustain non-stationary caused by the policy change of teammates. Prior works mainly concentrate on the policy change during the training phase or teammates altering cross episodes, ignoring the fact that teammates may suffer from policy change suddenly within an episode, which might lead to miscoordination and poor performance as a result. We formulate the problem as an open Dec-POMDP, where we control some agents to coordinate with uncontrolled teammates, whose policies could be changed within one episode. Then we develop a new framework, fast teammates adaptation (Fastap), to address the problem. Concretely, we first train versatile teammates' policies and assign them to different clusters via the Chinese Restaurant Process (CRP). Then, we train the controlled agent(s) to coordinate with the sampled uncontrolled teammates by capturing their identifications as context for fast adaptation. Finally, each agent applies its local information to anticipate the teammates' context for decision-making accordingly. This process proceeds alternately, leading to a robust policy that can adapt to any teammates during the decentralized execution phase. We show in multiple multi-agent benchmarks that Fastap can achieve superior performance than multiple baselines in stationary and non-stationary scenarios.

cs.MA↗
arXiv2023

Communication-Robust Multi-Agent Learning by Adaptable Auxiliary Multi-Agent Adversary Generation

Communication can promote coordination in cooperative Multi-Agent Reinforcement Learning (MARL). Nowadays, existing works mainly focus on improving the communication efficiency of agents, neglecting that real-world communication is much more challenging as there may exist noise or potential attackers. Thus the robustness of the communication-based policies becomes an emergent and severe issue that needs more exploration. In this paper, we posit that the ego system trained with auxiliary adversaries may handle this limitation and propose an adaptable method of Multi-Agent Auxiliary Adversaries Generation for robust Communication, dubbed MA3C, to obtain a robust communication-based policy. In specific, we introduce a novel message-attacking approach that models the learning of the auxiliary attacker as a cooperative problem under a shared goal to minimize the coordination ability of the ego system, with which every information channel may suffer from distinct message attacks. Furthermore, as naive adversarial training may impede the generalization ability of the ego system, we design an attacker population generation approach based on evolutionary learning. Finally, the ego system is paired with an attacker population and then alternatively trained against the continuously evolving attackers to improve its robustness, meaning that both the ego system and the attackers are adaptable. Extensive experiments on multiple benchmarks indicate that our proposed MA3C provides comparable or better robustness and generalization ability than other baselines.

cs.LG↗
arXiv2023

Multi-agent Continual Coordination via Progressive Task Contextualization

Cooperative Multi-agent Reinforcement Learning (MARL) has attracted significant attention and played the potential for many real-world applications. Previous arts mainly focus on facilitating the coordination ability from different aspects (e.g., non-stationarity, credit assignment) in single-task or multi-task scenarios, ignoring the stream of tasks that appear in a continual manner. This ignorance makes the continual coordination an unexplored territory, neither in problem formulation nor efficient algorithms designed. Towards tackling the mentioned issue, this paper proposes an approach Multi-Agent Continual Coordination via Progressive Task Contextualization, dubbed MACPro. The key point lies in obtaining a factorized policy, using shared feature extraction layers but separated independent task heads, each specializing in a specific class of tasks. The task heads can be progressively expanded based on the learned task contextualization. Moreover, to cater to the popular CTDE paradigm in MARL, each agent learns to predict and adopt the most relevant policy head based on local information in a decentralized manner. We show in multiple multi-agent benchmarks that existing continual learning methods fail, while MACPro is able to achieve close-to-optimal performance. More results also disclose the effectiveness of MACPro from multiple aspects like high generalization ability.

cs.MA↗
arXiv2023

Robust Multi-agent Communication via Multi-view Message Certification

Many multi-agent scenarios require message sharing among agents to promote coordination, hastening the robustness of multi-agent communication when policies are deployed in a message perturbation environment. Major relevant works tackle this issue under specific assumptions, like a limited number of message channels would sustain perturbations, limiting the efficiency in complex scenarios. In this paper, we take a further step addressing this issue by learning a robust multi-agent communication policy via multi-view message certification, dubbed CroMAC. Agents trained under CroMAC can obtain guaranteed lower bounds on state-action values to identify and choose the optimal action under a worst-case deviation when the received messages are perturbed. Concretely, we first model multi-agent communication as a multi-view problem, where every message stands for a view of the state. Then we extract a certificated joint message representation by a multi-view variational autoencoder (MVAE) that uses a product-of-experts inference network. For the optimization phase, we do perturbations in the latent space of the state for a certificate guarantee. Then the learned joint message representation is used to approximate the certificated state representation during training. Extensive experiments in several cooperative multi-agent benchmarks validate the effectiveness of the proposed CroMAC.

cs.MA↗
arXiv2023

Sim2Rec: A Simulator-based Decision-making Approach to Optimize Real-World Long-term User Engagement in Sequential Recommender Systems

Long-term user engagement (LTE) optimization in sequential recommender systems (SRS) is shown to be suited by reinforcement learning (RL) which finds a policy to maximize long-term rewards. Meanwhile, RL has its shortcomings, particularly requiring a large number of online samples for exploration, which is risky in real-world applications. One of the appealing ways to avoid the risk is to build a simulator and learn the optimal recommendation policy in the simulator. In LTE optimization, the simulator is to simulate multiple users' daily feedback for given recommendations. However, building a user simulator with no reality-gap, i.e., can predict user's feedback exactly, is unrealistic because the users' reaction patterns are complex and historical logs for each user are limited, which might mislead the simulator-based recommendation policy. In this paper, we present a practical simulator-based recommender policy training approach, Simulation-to-Recommendation (Sim2Rec) to handle the reality-gap problem for LTE optimization. Specifically, Sim2Rec introduces a simulator set to generate various possibilities of user behavior patterns, then trains an environment-parameter extractor to recognize users' behavior patterns in the simulators. Finally, a context-aware policy is trained to make the optimal decisions on all of the variants of the users based on the inferred environment-parameters. The policy is transferable to unseen environments (e.g., the real world) directly as it has learned to recognize all various user behavior patterns and to make the correct decisions based on the inferred environment-parameters. Experiments are conducted in synthetic environments and a real-world large-scale ride-hailing platform, DidiChuxing. The results show that Sim2Rec achieves significant performance improvement, and produces robust recommendations in unseen environments.

cs.IR↗
arXiv2023

Polarised parton distribution functions and proton spin

Supposing there exists an effective charge which defines an evolution scheme for both unpolarised and polarised parton distribution functions (DFs) that is all-orders exact and using Ansätze for hadron-scale proton polarised valence quark DFs, constrained by flavour-separated axial charges and insights from perturbative quantum chromodynamics, predictions are delivered for all proton polarised DFs at the scale $ζ_{\rm C}^2 = 3\,$GeV$^2$. The pointwise behaviour of the predicted DFs and, consequently, their moments, compare favourably with results inferred from data. Notably, flavour-separated singlet polarised DFs are small. On the other hand, the polarised gluon DF, $ΔG(x;ζ_{\rm C})$, is large and positive. Using our result, we predict $\int_{0.05}^1\,dx\,ΔG(x;ζ_{\rm C}) = 0.214(4)$ and that experimental measurements of the proton flavour-singlet axial charge should return $a_0^{\rm E}(ζ_{\rm C}) = 0.35(2)$.

hep-ph↗
arXiv2023

Knowledge is Power: Understanding Causality Makes Legal judgment Prediction Models More Generalizable and Robust

Legal Judgment Prediction (LJP), aiming to predict a judgment based on fact descriptions according to rule of law, serves as legal assistance to mitigate the great work burden of limited legal practitioners. Most existing methods apply various large-scale pre-trained language models (PLMs) finetuned in LJP tasks to obtain consistent improvements. However, we discover the fact that the state-of-the-art (SOTA) model makes judgment predictions according to irrelevant (or non-casual) information. The violation of rule of law not only weakens the robustness and generalization ability of models but also results in severe social problems like discrimination. In this paper, we use causal structural models (SCMs) to theoretically analyze how LJP models learn to make decisions and why they can succeed in passing the traditional testing paradigm without learning causality. According to our analysis, we provide two solutions intervening on data and model by causality, respectively. In detail, we first distinguish non-causal information by applying the open information extraction (OIE) technique. Then, we propose a method named the Causal Information Enhanced SAmpling Method (CIESAM) to eliminate the non-causal information from data. To validate our theoretical analysis, we further propose another method using our proposed Causality-Aware Self-Attention Mechanism (CASAM) to guide the model to learn the underlying causality knowledge in legal texts. The confidence of CASAM in learning causal information is higher than that of CIESAM. The extensive experimental results show that both our proposed methods achieve state-of-the-art (SOTA) performance on three commonly used legal-specific datasets. The stronger performance of CASAM further demonstrates that causality is the key to the robustness and generalization ability of models.

cs.CL↗
arXiv2023

The common trend of saltation particles on the surface of fast-rotating asteroids

An asteroid spun up to its critical limit has unique surface mechanical properties that its gravity and the centrifugal force largely balance, creating a relaxation environment where low-energy events such as mass shedding may trigger subsequent long complex motion of an asteroid's regolith grains. Exploring such an evolution process may provide key clues for understanding the early formation of multi-asteroid systems. This paper investigates the complex evolution process of loose particles becoming triggered by shedding events and the dependency of their dynamical propagation on the contact mechanical properties of the asteroid surface. We present a numerical model for tracking the trajectory of a shed particle that considers the collision between the particle and the surface of an asteroid. Monte Carlo simulations are performed to reflect the statistical behavior of shed particles. We also introduce zero-velocity surfaces to our data-based analysis in order to reveal the intrinsic invariance of the evolutionary processes. We used the average mechanical energy of the particle cloud to check the connection between contact property and the temporal-spatial distribution of the shed particles. We sketch a common evolutionary path of the particle in the vicinity of a fast-rotating asteroid, that is, particles dislodged from the unstable region will eventually enter, through several collisions with the surface, non-return orbits that launch from the minimum geopotential area of the unstable region. The common trend is independent of any particular asteroid morphology, and all shed particles enter the same evolutionary path. We also find that the orbital energy of the particle cloud is statistically independent of the surface contact property, meaning that the collision coefficient of restitution is a nonsensitive parameter in the outward spreading process of the shed particles.

astro-ph.EP↗
arXiv2023

Smart Contract Generation for Inter-Organizational Process Collaboration

Currently, inter-organizational process collaboration (IOPC) has been widely used in the design and development of distributed systems that support business process execution. Blockchain-based IOPC can establish trusted data sharing among participants, attracting more and more attention. The core of such study is to translate the graphical model (e.g., BPMN) into program code called smart contract that can be executed in the blockchain environment. In this context, a proper smart contract plays a vital role in the correct implementation of block-chain-based IOPC. In fact, the quality of graphical model affects the smart con-tract generation. Problematic models (e.g., deadlock) will result in incorrect contracts (causing unexpected behaviours). To avoid this undesired implementation, this paper explores to generate smart contracts by using the verified formal model as input instead of graphical model. Specifically, we introduce a prototype framework that supports the automatic generation of smart contracts, providing an end-to-end solution from modeling, verification, translation to implementation. One of the cores of this framework is to provide a CSP#-based formalization for the BPMN collaboration model from the perspective of message interaction. This formalization provides precise execution semantics and model verification for graphical models, and a verified formal model for smart contract generation. Another novelty is that it introduces a syntax tree-based translation algorithm to directly map the formal model into a smart contract. The required formalism, verification and translation techniques are transparent to users without imposing additional burdens. Finally, a set of experiments shows the effectiveness of the framework.

cs.SE↗
arXiv2023

Beware of Instantaneous Dependence in Reinforcement Learning

Playing an important role in Model-Based Reinforcement Learning (MBRL), environment models aim to predict future states based on the past. Existing works usually ignore instantaneous dependence in the state, that is, assuming that the future state variables are conditionally independent given the past states. However, instantaneous dependence is prevalent in many RL environments. For instance, in the stock market, instantaneous dependence can exist between two stocks because the fluctuation of one stock can quickly affect the other and the resolution of price change is lower than that of the effect. In this paper, we prove that with few exceptions, ignoring instantaneous dependence can result in suboptimal policy learning in MBRL. To address the suboptimality problem, we propose a simple plug-and-play method to enable existing MBRL algorithms to take instantaneous dependence into account. Through experiments on two benchmarks, we (1) confirm the existence of instantaneous dependence with visualization; (2) validate our theoretical findings that ignoring instantaneous dependence leads to suboptimal policy; (3) verify that our method effectively enables reinforcement learning with instantaneous dependence and improves policy performance.

cs.LG↗
arXiv2023

Monocular Camera Mapping with Pose-Guided Optimization: Enhancing Marking-Level HD Map Accuracy

Marking-level high-definition maps (HD maps) are of great significance for autonomous vehicles (AVs), especially in large-scale, appearance-changing scenarios where AVs rely on markings for localization and lanes for safe driving. In this paper, we propose a pose-guided optimization framework for automatically building a marking-level HD map with accurate markings positions using a simple sensor setup (one or more monocular cameras). We optimize the position of the marking corners to fit the result of marking segmentation and simultaneously optimize the inverse perspective mapping (IPM) matrix of the corresponding camera to obtain an accurate transformation from the front view image to the bird's-eye view (BEV). In the quantitative evaluation, the built HD map almost attains centimeter-level accuracy. The accuracy of the optimized IPM matrix is similar to that of the manual calibration. The method can also be generalized to build HD maps in a broader sense by increasing the types of recognizable markings. The supplementary materials and videos are available at http://liuhongji.site/V2HDM-Mono/.

cs.RO↗
arXiv2023

How To Guide Your Learner: Imitation Learning with Active Adaptive Expert Involvement

Imitation learning aims to mimic the behavior of experts without explicit reward signals. Passive imitation learning methods which use static expert datasets typically suffer from compounding error, low sample efficiency, and high hyper-parameter sensitivity. In contrast, active imitation learning methods solicit expert interventions to address the limitations. However, recent active imitation learning methods are designed based on human intuitions or empirical experience without theoretical guarantee. In this paper, we propose a novel active imitation learning framework based on a teacher-student interaction model, in which the teacher's goal is to identify the best teaching behavior and actively affect the student's learning process. By solving the optimization objective of this framework, we propose a practical implementation, naming it AdapMen. Theoretical analysis shows that AdapMen can improve the error bound and avoid compounding error under mild conditions. Experiments on the MetaDrive benchmark and Atari 2600 games validate our theoretical analysis and show that our method achieves near-expert performance with much less expert involvement and total sampling steps than previous methods. The code is available at https://github.com/liuxhym/AdapMen.

cs.LG↗
arXiv2023

A Partially Functional Linear Modeling Framework for Integrating Genetic, Imaging, and Clinical Data

This paper is motivated by the joint analysis of genetic, imaging, and clinical (GIC) data collected in the Alzheimer's Disease Neuroimaging Initiative (ADNI) study. We propose a regression framework based on partially functional linear regression models to map high-dimensional GIC-related pathways for Alzheimer's Disease (AD). We develop a joint model selection and estimation procedure by embedding imaging data in the reproducing kernel Hilbert space and imposing the L0 penalty for the coefficients of genetic variables. We apply the proposed method to the ADNI dataset to identify important features from tens of thousands of genetic polymorphisms (reduced from millions using a preprocessing step) and study the effects of a certain set of informative genetic variants and the baseline hippocampus surface on thirteen future cognitive scores measuring different aspects of cognitive function. We explore the shared and different heritability patterns of these cognitive scores. Analysis results suggest that both the hippocampal and genetic data have heterogeneous effects on different scores, with the trend that the value of both hippocampi is negatively associated with the severity of cognition deficits. Polygenic effects are observed for all thirteen cognitive scores. The well-known APOE4 genotype only explains a small part of cognitive function. Shared genetic etiology exists, however, greater genetic heterogeneity exists within disease classifications after accounting for the baseline diagnosis status. These analyses are useful in further investigation of functional mechanisms for AD evolution.

stat.ME↗
arXiv2023

Efficient Communication via Self-supervised Information Aggregation for Online and Offline Multi-agent Reinforcement Learning

Utilizing messages from teammates can improve coordination in cooperative Multi-agent Reinforcement Learning (MARL). Previous works typically combine raw messages of teammates with local information as inputs for policy. However, neglecting message aggregation poses significant inefficiency for policy learning. Motivated by recent advances in representation learning, we argue that efficient message aggregation is essential for good coordination in cooperative MARL. In this paper, we propose Multi-Agent communication via Self-supervised Information Aggregation (MASIA), where agents can aggregate the received messages into compact representations with high relevance to augment the local policy. Specifically, we design a permutation invariant message encoder to generate common information-aggregated representation from messages and optimize it via reconstructing and shooting future information in a self-supervised manner. Hence, each agent would utilize the most relevant parts of the aggregated representation for decision-making by a novel message extraction mechanism. Furthermore, considering the potential of offline learning for real-world applications, we build offline benchmarks for multi-agent communication, which is the first as we know. Empirical results demonstrate the superiority of our method in both online and offline settings. We also release the built offline benchmarks in this paper as a testbed for communication ability validation to facilitate further future research.

cs.LG↗
arXiv2023

Natural Language-conditioned Reinforcement Learning with Inside-out Task Language Development and Translation

Natural Language-conditioned reinforcement learning (RL) enables the agents to follow human instructions. Previous approaches generally implemented language-conditioned RL by providing human instructions in natural language (NL) and training a following policy. In this outside-in approach, the policy needs to comprehend the NL and manage the task simultaneously. However, the unbounded NL examples often bring much extra complexity for solving concrete RL tasks, which can distract policy learning from completing the task. To ease the learning burden of the policy, we investigate an inside-out scheme for natural language-conditioned RL by developing a task language (TL) that is task-related and unique. The TL is used in RL to achieve highly efficient and effective policy training. Besides, a translator is trained to translate NL into TL. We implement this scheme as TALAR (TAsk Language with predicAte Representation) that learns multiple predicates to model object relationships as the TL. Experiments indicate that TALAR not only better comprehends NL instructions but also leads to a better instruction-following policy that improves 13.4% success rate and adapts to unseen expressions of NL instruction. The TL can also be an effective task abstraction, naturally compatible with hierarchical RL.

cs.CL↗
arXiv2023

Real-time Blind Deblurring Based on Lightweight Deep-Wiener-Network

In this paper, we address the problem of blind deblurring with high efficiency. We propose a set of lightweight deep-wiener-network to finish the task with real-time speed. The Network contains a deep neural network for estimating parameters of wiener networks and a wiener network for deblurring. Experimental evaluations show that our approaches have an edge on State of the Art in terms of inference times and numbers of parameters. Two of our models can reach a speed of 100 images per second, which is qualified for real-time deblurring. Further research may focus on some real-world applications of deblurring with our models.

cs.CV↗