arXiv subjects

He Bai

Publications and source records attributed to He Bai.

At least 37 records · Page 2Linked to original sources

arXiv2024

Distributed Multi-Agent Reinforcement Learning Based on Graph-Induced Local Value Functions

Achieving distributed reinforcement learning (RL) for large-scale cooperative multi-agent systems (MASs) is challenging because: (i) each agent has access to only limited information; (ii) issues on convergence or computational complexity emerge due to the curse of dimensionality. In this paper, we propose a general computationally efficient distributed framework for cooperative multi-agent reinforcement learning (MARL) by utilizing the structures of graphs involved in this problem. We introduce three coupling graphs describing three types of inter-agent couplings in MARL, namely, the state graph, the observation graph and the reward graph. By further considering a communication graph, we propose two distributed RL approaches based on local value-functions derived from the coupling graphs. The first approach is able to reduce sample complexity significantly under specific conditions on the aforementioned four graphs. The second approach provides an approximate solution and can be efficient even for problems with dense coupling graphs. Here there is a trade-off between minimizing the approximation error and reducing the computational complexity. Simulations show that our RL algorithms have a significantly improved scalability to large-scale MASs compared with centralized and consensus-based distributed RL algorithms.

cs.LG↗

arXiv2024

Rephrasing the Web: A Recipe for Compute and Data-Efficient Language Modeling

Large language models are trained on massive scrapes of the web, which are often unstructured, noisy, and poorly phrased. Current scaling laws show that learning from such data requires an abundance of both compute and data, which grows with the size of the model being trained. This is infeasible both because of the large compute costs and duration associated with pre-training, and the impending scarcity of high-quality data on the web. In this work, we propose Web Rephrase Augmented Pre-training ($\textbf{WRAP}$) that uses an off-the-shelf instruction-tuned model prompted to paraphrase documents on the web in specific styles such as "like Wikipedia" or in "question-answer format" to jointly pre-train LLMs on real and synthetic rephrases. First, we show that using WRAP on the C4 dataset, which is naturally noisy, speeds up pre-training by $\sim3x$. At the same pre-training compute budget, it improves perplexity by more than 10% on average across different subsets of the Pile, and improves zero-shot question answer accuracy across 13 tasks by more than 2%. Second, we investigate the impact of the re-phrasing style on the performance of the model, offering insights into how the composition of the training data can impact the performance of LLMs in OOD settings. Our gains are attributed to the fact that re-phrased synthetic data has higher utility than just real data because it (i) incorporates style diversity that closely reflects downstream evaluation style, and (ii) has higher 'quality' than web-scraped data.

cs.CL↗

arXiv2024

Asynchronous Local Computations in Distributed Bayesian Learning

Due to the expanding scope of machine learning (ML) to the fields of sensor networking, cooperative robotics and many other multi-agent systems, distributed deployment of inference algorithms has received a lot of attention. These algorithms involve collaboratively learning unknown parameters from dispersed data collected by multiple agents. There are two competing aspects in such algorithms, namely, intra-agent computation and inter-agent communication. Traditionally, algorithms are designed to perform both synchronously. However, certain circumstances need frugal use of communication channels as they are either unreliable, time-consuming, or resource-expensive. In this paper, we propose gossip-based asynchronous communication to leverage fast computations and reduce communication overhead simultaneously. We analyze the effects of multiple (local) intra-agent computations by the active agents between successive inter-agent communications. For local computations, Bayesian sampling via unadjusted Langevin algorithm (ULA) MCMC is utilized. The communication is assumed to be over a connected graph (e.g., as in decentralized learning), however, the results can be extended to coordinated communication where there is a central server (e.g., federated learning). We theoretically quantify the convergence rates in the process. To demonstrate the efficacy of the proposed algorithm, we present simulations on a toy problem as well as on real world data sets to train ML models to perform classification tasks. We observe faster initial convergence and improved performance accuracy, especially in the low data range. We achieve on average 78% and over 90% classification accuracy respectively on the Gamma Telescope and mHealth data sets from the UCI ML repository.

cs.LG↗

arXiv2023

Construction of Paired Knowledge Graph-Text Datasets Informed by Cyclic Evaluation

Datasets that pair Knowledge Graphs (KG) and text together (KG-T) can be used to train forward and reverse neural models that generate text from KG and vice versa. However models trained on datasets where KG and text pairs are not equivalent can suffer from more hallucination and poorer recall. In this paper, we verify this empirically by generating datasets with different levels of noise and find that noisier datasets do indeed lead to more hallucination. We argue that the ability of forward and reverse models trained on a dataset to cyclically regenerate source KG or text is a proxy for the equivalence between the KG and the text in the dataset. Using cyclic evaluation we find that manually created WebNLG is much better than automatically created TeKGen and T-REx. Guided by these observations, we construct a new, improved dataset called LAGRANGE using heuristics meant to improve equivalence between KG and text and show the impact of each of the heuristics on cyclic evaluation. We also construct two synthetic datasets using large language models (LLMs), and observe that these are conducive to models that perform significantly well on cyclic generation of text, but less so on cyclic generation of KGs, probably because of a lack of a consistent underlying ontology.

cs.CL↗

arXiv2023

A Study on Workload Assessment and Usability of Wind-Aware User Interface for Small Unmanned Aircraft System Remote Operations

This study evaluates pilots' cognitive workload and situational awareness during remote small unmanned aircraft system operations in different wind conditions. To complement the urban air mobility concept that envisions safe, sustainable, and accessible air transportation, we conduct multiple experiments in a realistic wind-aware simulator-user interface pipeline. Experiments are performed with basic and wind-aware displays in several wind conditions to assess how complex wind fields impact pilots' cognitive resources. Post-hoc analysis reveals that providing pilots with real-time wind information improves situational awareness while decreasing cognitive workload.

cs.HC↗

arXiv2023

Online Observer-Based Inverse Reinforcement Learning

In this paper, a novel approach to the output-feedback inverse reinforcement learning (IRL) problem is developed by casting the IRL problem, for linear systems with quadratic cost functions, as a state estimation problem. Two observer-based techniques for IRL are developed, including a novel observer method that re-uses previous state estimates via history stacks. Theoretical guarantees for convergence and robustness are established under appropriate excitation conditions. Simulations demonstrate the performance of the developed observers and filters under noisy and noise-free measurements.

eess.SY↗

arXiv2022

ERNIE-SAT: Speech and Text Joint Pretraining for Cross-Lingual Multi-Speaker Text-to-Speech

Speech representation learning has improved both speech understanding and speech synthesis tasks for single language. However, its ability in cross-lingual scenarios has not been explored. In this paper, we extend the pretraining method for cross-lingual multi-speaker speech synthesis tasks, including cross-lingual multi-speaker voice cloning and cross-lingual multi-speaker speech editing. We propose a speech-text joint pretraining framework, where we randomly mask the spectrogram and the phonemes given a speech example and its transcription. By learning to reconstruct the masked parts of the input in different languages, our model shows great improvements over speaker-embedding-based multi-speaker TTS methods. Moreover, our framework is end-to-end for both the training and the inference without any finetuning effort. In cross-lingual multi-speaker voice cloning and cross-lingual multi-speaker speech editing tasks, our experiments show that our model outperforms speaker-embedding-based multi-speaker TTS methods.

eess.AS↗

arXiv2022

Asynchronous Bayesian Learning over a Network

We present a practical asynchronous data fusion model for networked agents to perform distributed Bayesian learning without sharing raw data. Our algorithm uses a gossip-based approach where pairs of randomly selected agents employ unadjusted Langevin dynamics for parameter sampling. We also introduce an event-triggered mechanism to further reduce communication between gossiping agents. These mechanisms drastically reduce communication overhead and help avoid bottlenecks commonly experienced with distributed algorithms. In addition, the reduced link utilization by the algorithm is expected to increase resiliency to occasional link failure. We establish mathematical guarantees for our algorithm and demonstrate its effectiveness via numerical experiments.

cs.LG↗

arXiv2022

XRICL: Cross-lingual Retrieval-Augmented In-Context Learning for Cross-lingual Text-to-SQL Semantic Parsing

In-context learning using large language models has recently shown surprising results for semantic parsing tasks such as Text-to-SQL translation. Prompting GPT-3 or Codex using several examples of question-SQL pairs can produce excellent results, comparable to state-of-the-art finetuning-based models. However, existing work primarily focuses on English datasets, and it is unknown whether large language models can serve as competitive semantic parsers for other languages. To bridge this gap, our work focuses on cross-lingual Text-to-SQL semantic parsing for translating non-English utterances into SQL queries based on an English schema. We consider a zero-shot transfer learning setting with the assumption that we do not have any labeled examples in the target language (but have annotated examples in English). This work introduces the XRICL framework, which learns to retrieve relevant English exemplars for a given query to construct prompts. We also include global translation exemplars for a target language to facilitate the translation process for large language models. To systematically evaluate our model, we construct two new benchmark datasets, XSpider and XKaggle-dbqa, which include questions in Chinese, Vietnamese, Farsi, and Hindi. Our experiments show that XRICL effectively leverages large pre-trained language models to outperform existing baselines. Data and code are publicly available at https://github.com/Impavidity/XRICL.

cs.CL↗

arXiv2022

Emergent magnetic states and tunable exchange bias at all 3d nitride heterointerfaces

Interfacial magnetism stimulates the discovery of giant magnetoresistance and spin-orbital coupling across the heterointerfaces, facilitating the intimate correlation between spin transport and complex magnetic structures. Over decades, functional heterointerfaces composed of nitrides are seldomly explored due to the difficulty in synthesizing high-quality and correct composition nitride films. Here we report the fabrication of single-crystalline ferromagnetic Fe3N thin films with precisely controlled thickness. As film thickness decreasing, the magnetization deteriorates dramatically, and electronic state transits from metallic to insulating. Strikingly, the high-temperature ferromagnetism maintains in a Fe3N layer with a thickness down to 2 u. c. (~ 8 Å). The magnetoresistance exhibits a strong in-plane anisotropy and meanwhile the anomalous Hall resistance reserves its sign when Fe3N layer thickness exceeds 5 u. c. Furthermore, we observe a sizable exchange bias at the interfaces between a ferromagnetic Fe3N and an antiferromagnetic CrN. The exchange bias field and saturation moment strongly depend on the controllable bending curvature using cylinder diameter engineering (CDE) technique, implying the tunable magnetic states under lattice deformation. This work provides a guideline for exploring functional nitride films and applying their interfacial phenomena for innovative perspectives towards the practical applications.

cond-mat.mtrl-sci↗

arXiv2022

Braiding lateral morphotropic grain boundary in homogeneitic oxides

Interfaces formed by correlated oxides offer a critical avenue for discovering emergent phenomena and quantum states. However, the fabrication of oxide interfaces with variable crystallographic orientations and strain states integrated along a film plane is extremely challenge by conventional layer-by-layer stacking or self-assembling. Here, we report the creation of morphotropic grain boundaries (GBs) in laterally interconnected cobaltite homostructures. Single-crystalline substrates and suspended ultrathin freestanding membranes provide independent templates for coherent epitaxy and constraint on the growth orientation, resulting in seamless and atomically sharp GBs. Electronic states and magnetic behavior in hybrid structures are laterally modulated and isolated by GBs, enabling artificially engineered functionalities in the planar matrix. Our work offers a simple and scalable method for fabricating unprecedented innovative interfaces through controlled synthesis routes as well as provides a platform for exploring potential applications in neuromorphics, solid state batteries, and catalysis.

cond-mat.mtrl-sci↗

arXiv2022

A$^3$T: Alignment-Aware Acoustic and Text Pretraining for Speech Synthesis and Editing

Recently, speech representation learning has improved many speech-related tasks such as speech recognition, speech classification, and speech-to-text translation. However, all the above tasks are in the direction of speech understanding, but for the inverse direction, speech synthesis, the potential of representation learning is yet to be realized, due to the challenging nature of generating high-quality speech. To address this problem, we propose our framework, Alignment-Aware Acoustic-Text Pretraining (A$^3$T), which reconstructs masked acoustic signals with text input and acoustic-text alignment during training. In this way, the pretrained model can generate high quality reconstructed spectrogram, which can be applied to the speech editing and unseen speaker TTS directly. Experiments show A$^3$T outperforms SOTA models on speech editing, and improves multi-speaker speech synthesis without the external speaker verification model.

eess.AS↗

arXiv2022

Better Language Model with Hypernym Class Prediction

Class-based language models (LMs) have been long devised to address context sparsity in $n$-gram LMs. In this study, we revisit this approach in the context of neural LMs. We hypothesize that class-based prediction leads to an implicit context aggregation for similar words and thus can improve generalization for rare words. We map words that have a common WordNet hypernym to the same class and train large neural LMs by gradually annealing from predicting the class to token prediction during training. Empirically, this curriculum learning strategy consistently improves perplexity over various large, highly-performant state-of-the-art Transformer-based models on two datasets, WikiText-103 and Arxiv. Our analysis shows that the performance improvement is achieved without sacrificing performance on rare words. Finally, we document other attempts that failed to yield empirical gains, and discuss future directions for the adoption of class-based LMs on a larger scale.

cs.CL↗

arXiv2022

Quantifying the Dzyaloshinskii-Moriya Interaction Induced by the Bulk Magnetic Asymmetry

A broken interfacial inversion symmetry in ultrathin ferromagnet/heavy metal (FM/HM) bilayers is generally believed to be a prerequisite for accommodating the Dzyaloshinskii-Moriya interaction (DMI) and for stabilizing chiral spin textures. In these bilayers, the strength of the DMI decays as the thickness of the FM layer increases and vanishes around a few nanometers. In the present study, through synthesizing relatively thick films of compositions CoPt or FePt, CoCu or FeCu, FeGd and FeNi, contributions to DMI from the composition gradient induced bulk magnetic asymmetry (BMA) and spin-orbit coupling (SOC) are systematically examined. Using Brillouin light scattering spectroscopy, both the sign and amplitude of DMI in films with controllable direction and strength of BMA, in the presence and absence of SOC are experimentally studied. In particular, we show that a sizable amplitude of DMI (0.15 mJ/m^2) can be realized in CoPt or FePt films with BMA and strong SOC, whereas negligible DMI strengths are observed in other thick films with BMA but without significant SOC. The pivotal roles of BMA and SOC are further examined based on the three-site Fert-Levy model and first-principles calculations. It is expected that our findings may help to further understand the origin of chiral magnetism and to design novel non-collinear spin textures.

cond-mat.mtrl-sci↗

arXiv2022

Distributed Cooperative Multi-Agent Reinforcement Learning with Directed Coordination Graph

Existing distributed cooperative multi-agent reinforcement learning (MARL) frameworks usually assume undirected coordination graphs and communication graphs while estimating a global reward via consensus algorithms for policy evaluation. Such a framework may induce expensive communication costs and exhibit poor scalability due to requirement of global consensus. In this work, we study MARLs with directed coordination graphs, and propose a distributed RL algorithm where the local policy evaluations are based on local value functions. The local value function of each agent is obtained by local communication with its neighbors through a directed learning-induced communication graph, without using any consensus algorithm. A zeroth-order optimization (ZOO) approach based on parameter perturbation is employed to achieve gradient estimation. By comparing with existing ZOO-based RL algorithms, we show that our proposed distributed RL algorithm guarantees high scalability. A distributed resource allocation example is shown to illustrate the effectiveness of our algorithm.

cs.MA↗

arXiv2021

Room-temperature ferromagnetism at an oxide/nitride interface

Heterointerfaces have led to the discovery of novel electronic and magnetic states because of their strongly entangled electronic degrees of freedom. Single-phase chromium compounds always exhibit antiferromagnetism following the prediction of Goodenough-Kanamori rules. So far, exchange coupling between chromium ions via hetero-anions has not been explored and the associated quantum states is unknown. Here we report the successful epitaxial synthesis and characterizations of chromium oxide (Cr2O3)-chromium nitride (CrN) superlattices. Room-temperature ferromagnetic spin ordering is achieved at the interfaces between these two antiferromagnets, and the magnitude of the effect decays with increasing layer thickness. First-principles calculations indicate that robust ferromagnetic spin interaction between Cr3+ ions via anion-hybridizations across the interface yields the lowest total energy. This work opens the door to fundamental understanding of the unexpected and exceptional properties of oxide-nitride interfaces and provides access to hidden phases at low-dimensional quantum heterostructures.

cond-mat.mtrl-sci↗

arXiv2021

Cross-Lingual Training with Dense Retrieval for Document Retrieval

Dense retrieval has shown great success in passage ranking in English. However, its effectiveness in document retrieval for non-English languages remains unexplored due to the limitation in training resources. In this work, we explore different transfer techniques for document ranking from English annotations to multiple non-English languages. Our experiments on the test collections in six languages (Chinese, Arabic, French, Hindi, Bengali, Spanish) from diverse language families reveal that zero-shot model-based transfer using mBERT improves the search quality in non-English mono-lingual retrieval. Also, we find that weakly-supervised target language transfer yields competitive performances against the generation-based target language transfer that requires external translators and query generators.

cs.CL↗

arXiv2021

Variance Reduction of Quadcopter Trajectory Tracking in Turbulent Wind

We consider a quadcopter operating in a turbulent windy environment. The turbulent environment may be imposed on a quadcopter by structures, landscapes, terrains and most importantly by the unique physical phenomena in the lower atmosphere. Turbulence can negatively impact quadcopter's performance and operations. Modeling turbulence as a stochastic random input, we investigate control designs that can reduce the turbulence effects on the quadcopter's motion. In particular, we design a minimum cost variance (MCV) controller aiming to minimize the cost in terms of its weighted sum of mean and variance. We linearize the quadcopter dynamics and examine the MCV controller derived from a set of coupled algebraic Riccati equations (CARE) with full-state feedback. Our preliminary simulation results show reduction in variance and in mean trajectory tracking error compared to a traditional linear quadratic regulator (LQR).

eess.SY↗