arXiv subjects

Peng Li

Publications and source records attributed to Peng Li.

At least 343 records · Page 19Linked to original sources

arXiv2022

Manual-Guided Dialogue for Flexible Conversational Agents

How to build and use dialogue data efficiently, and how to deploy models in different domains at scale can be two critical issues in building a task-oriented dialogue system. In this paper, we propose a novel manual-guided dialogue scheme to alleviate these problems, where the agent learns the tasks from both dialogue and manuals. The manual is an unstructured textual document that guides the agent in interacting with users and the database during the conversation. Our proposed scheme reduces the dependence of dialogue models on fine-grained domain ontology, and makes them more flexible to adapt to various domains. We then contribute a fully-annotated multi-domain dataset MagDial to support our scheme. It introduces three dialogue modeling subtasks: instruction matching, argument filling, and response generation. Modeling these subtasks is consistent with the human agent's behavior patterns. Experiments demonstrate that the manual-guided dialogue scheme improves data efficiency and domain scalability in building dialogue systems. The dataset and benchmark will be publicly available for promoting future research.

cs.CL↗

arXiv2022

A Closed-Loop Perception, Decision-Making and Reasoning Mechanism for Human-Like Navigation

Reliable navigation systems have a wide range of applications in robotics and autonomous driving. Current approaches employ an open-loop process that converts sensor inputs directly into actions. However, these open-loop schemes are challenging to handle complex and dynamic real-world scenarios due to their poor generalization. Imitating human navigation, we add a reasoning process to convert actions back to internal latent states, forming a two-stage closed loop of perception, decision-making, and reasoning. Firstly, VAE-Enhanced Demonstration Learning endows the model with the understanding of basic navigation rules. Then, two dual processes in RL-Enhanced Interaction Learning generate reward feedback for each other and collectively enhance obstacle avoidance capability. The reasoning model can substantially promote generalization and robustness, and facilitate the deployment of the algorithm to real-world robots without elaborate transfers. Experiments show our method is more adaptable to novel scenarios compared with state-of-the-art approaches.

cs.RO↗

arXiv2022

LF-VIO: A Visual-Inertial-Odometry Framework for Large Field-of-View Cameras with Negative Plane

Visual-inertial-odometry has attracted extensive attention in the field of autonomous driving and robotics. The size of Field of View (FoV) plays an important role in Visual-Odometry (VO) and Visual-Inertial-Odometry (VIO), as a large FoV enables to perceive a wide range of surrounding scene elements and features. However, when the field of the camera reaches the negative half plane, one cannot simply use [u,v,1]^T to represent the image feature points anymore. To tackle this issue, we propose LF-VIO, a real-time VIO framework for cameras with extremely large FoV. We leverage a three-dimensional vector with unit length to represent feature points, and design a series of algorithms to overcome this challenge. To address the scarcity of panoramic visual odometry datasets with ground-truth location and pose, we present the PALVIO dataset, collected with a Panoramic Annular Lens (PAL) system with an entire FoV of 360°x(40°-120°) and an IMU sensor. With a comprehensive variety of experiments, the proposed LF-VIO is verified on both the established PALVIO benchmark and a public fisheye camera dataset with a FoV of 360°x(0°-93.5°). LF-VIO outperforms state-of-the-art visual-inertial-odometry methods. Our dataset and code are made publicly available at https://github.com/flysoaryun/LF-VIO

cs.CV↗

arXiv2022

ELLE: Efficient Lifelong Pre-training for Emerging Data

Current pre-trained language models (PLM) are typically trained with static data, ignoring that in real-world scenarios, streaming data of various sources may continuously grow. This requires PLMs to integrate the information from all the sources in a lifelong manner. Although this goal could be achieved by exhaustive pre-training on all the existing data, such a process is known to be computationally expensive. To this end, we propose ELLE, aiming at efficient lifelong pre-training for emerging data. Specifically, ELLE consists of (1) function preserved model expansion, which flexibly expands an existing PLM's width and depth to improve the efficiency of knowledge acquisition; and (2) pre-trained domain prompts, which disentangle the versatile knowledge learned during pre-training and stimulate the proper knowledge for downstream tasks. We experiment ELLE with streaming data from 5 domains on BERT and GPT. The results show the superiority of ELLE over various lifelong learning baselines in both pre-training efficiency and downstream performances. The codes are publicly available at https://github.com/thunlp/ELLE.

cs.CL↗

arXiv2022

Evidence of Magnon-Mediated Orbital Magnetism in a Quasi-2D Topological Magnon Insulator

We explore spin dynamics in Cu(1,3-bdc), a quasi-2D topological magnon insulator. The results show that the thermal evolution of Landé $g$-factor ($g$) is anisotropic: $g_\textrm{in-plane}$ reduces while $g_\textrm{out-plane}$ increases with increasing temperature $T$. Moreover, the anisotropy of the $g$-factor ($Δg$) and the anisotropy of saturation magnetization ($ΔM_\textrm{s}$) are correlated below 4 K, but they diverge above 4 K. We show that the electronic orbital moment contributes to the $g$ anisotropy at lower $T$, while the topological orbital moment induced by thermally excited spin chirality dictates the $g$ anisotropy at higher $T$. Our work suggests an interplay among topology, spin chirality, and orbital magnetism in Cu(1,3-bdc).

physics.app-ph↗

arXiv2022

CF-YOLO: Cross Fusion YOLO for Object Detection in Adverse Weather with a High-quality Real Snow Dataset

Snow is one of the toughest adverse weather conditions for object detection (OD). Currently, not only there is a lack of snowy OD datasets to train cutting-edge detectors, but also these detectors have difficulties learning latent information beneficial for detection in snow. To alleviate the two above problems, we first establish a real-world snowy OD dataset, named RSOD. Besides, we develop an unsupervised training strategy with a distinctive activation function, called $Peak \ Act$, to quantitatively evaluate the effect of snow on each object. Peak Act helps grading the images in RSOD into four-difficulty levels. To our knowledge, RSOD is the first quantitatively evaluated and graded snowy OD dataset. Then, we propose a novel Cross Fusion (CF) block to construct a lightweight OD network based on YOLOv5s (call CF-YOLO). CF is a plug-and-play feature aggregation module, which integrates the advantages of Feature Pyramid Network and Path Aggregation Network in a simpler yet more flexible form. Both RSOD and CF lead our CF-YOLO to possess an optimization ability for OD in real-world snow. That is, CF-YOLO can handle unfavorable detection problems of vagueness, distortion and covering of snow. Experiments show that our CF-YOLO achieves better detection results on RSOD, compared to SOTAs. The code and dataset are available at https://github.com/qqding77/CF-YOLO-and-RSOD.

cs.CV↗

arXiv2022

GoG: Relation-aware Graph-over-Graph Network for Visual Dialog

Visual dialog, which aims to hold a meaningful conversation with humans about a given image, is a challenging task that requires models to reason the complex dependencies among visual content, dialog history, and current questions. Graph neural networks are recently applied to model the implicit relations between objects in an image or dialog. However, they neglect the importance of 1) coreference relations among dialog history and dependency relations between words for the question representation; and 2) the representation of the image based on the fully represented question. Therefore, we propose a novel relation-aware graph-over-graph network (GoG) for visual dialog. Specifically, GoG consists of three sequential graphs: 1) H-Graph, which aims to capture coreference relations among dialog history; 2) History-aware Q-Graph, which aims to fully understand the question through capturing dependency relations between words based on coreference resolution on the dialog history; and 3) Question-aware I-Graph, which aims to capture the relations between objects in an image based on fully question representation. As an additional feature representation module, we add GoG to the existing visual dialogue model. Experimental results show that our model outperforms the strong baseline in both generative and discriminative settings by a significant margin.

cs.CL↗

arXiv2022

A Computational Framework of Cortical Microcircuits Approximates Sign-concordant Random Backpropagation

Several recent studies attempt to address the biological implausibility of the well-known backpropagation (BP) method. While promising methods such as feedback alignment, direct feedback alignment, and their variants like sign-concordant feedback alignment tackle BP's weight transport problem, their validity remains controversial owing to a set of other unsolved issues. In this work, we answer the question of whether it is possible to realize random backpropagation solely based on mechanisms observed in neuroscience. We propose a hypothetical framework consisting of a new microcircuit architecture and its supporting Hebbian learning rules. Comprising three types of cells and two types of synaptic connectivity, the proposed microcircuit architecture computes and propagates error signals through local feedback connections and supports the training of multi-layered spiking neural networks with a globally defined spiking error function. We employ the Hebbian rule operating in local compartments to update synaptic weights and achieve supervised learning in a biologically plausible manner. Finally, we interpret the proposed framework from an optimization point of view and show its equivalence to sign-concordant feedback alignment. The proposed framework is benchmarked on several datasets including MNIST and CIFAR10, demonstrating promising BP-comparable accuracy.

cs.NE↗

arXiv2022

A Simple but Effective Pluggable Entity Lookup Table for Pre-trained Language Models

Pre-trained language models (PLMs) cannot well recall rich factual knowledge of entities exhibited in large-scale corpora, especially those rare entities. In this paper, we propose to build a simple but effective Pluggable Entity Lookup Table (PELT) on demand by aggregating the entity's output representations of multiple occurrences in the corpora. PELT can be compatibly plugged as inputs to infuse supplemental entity knowledge into PLMs. Compared to previous knowledge-enhanced PLMs, PELT only requires 0.2%-5% pre-computation with capability of acquiring knowledge from out-of-domain corpora for domain adaptation scenario. The experiments on knowledge-related tasks demonstrate that our method, PELT, can flexibly and effectively transfer entity knowledge from related corpora into PLMs with different architectures.

cs.CL↗

arXiv2022

Anisotropic Dzyaloshinskii-Moriya interaction and topological magnetism in two-dimensional magnets protected by P4-m2 crystal symmetry

As a fundamental magnetic parameter, Dzyaloshinskii-Moriya interaction (DMI), has gained a great deal of attention in the last two decades due to its critical role in formation of magnetic skyrmions. Recent discoveries of two-dimensional (2D) van der Waals (vdW) magnets has also gained a great deal of attention due to appealing physical properties, such as gate tunability, flexibility and miniaturization. Intensive studies have shown that isotropic DMI stabilizes ferromagnetic (FM) topological spin textures in 2D magnets or their corresponding heterostructures. However, the investigation of anisotropic DMI and antiferromagnetic (AFM) topological spin configurations remains elusive. Here, we propose and demonstrate that a family of 2D magnets with P4-m2 symmetry-protected anisotropic DMI. More interestingly, various topological spin configurations, including FM/AFM antiskyrmion and AFM vortex-antivortex pair, emerge in this family. These results give a general method to design anisotropic DMI and pave the way towards topological magnetism in 2D materials using crystal symmetry.

cond-mat.mtrl-sci↗

arXiv2022

Rethinking the Promotion Brought by Contrastive Learning to Semi-Supervised Node Classification

Graph Contrastive Learning (GCL) has proven highly effective in promoting the performance of Semi-Supervised Node Classification (SSNC). However, existing GCL methods are generally transferred from other fields like CV or NLP, whose underlying working mechanism remains under-explored. In this work, we first deeply probe the working mechanism of GCL in SSNC, and find that the promotion brought by GCL is severely unevenly distributed: the improvement mainly comes from subgraphs with less annotated information, which is fundamentally different from contrastive learning in other fields. However, existing GCL methods generally ignore this uneven distribution of annotated information and apply GCL evenly to the whole graph. To remedy this issue and further improve GCL in SSNC, we propose the Topology InFormation gain-Aware Graph Contrastive Learning (TIFA-GCL) framework that considers the annotated information distribution across graph in GCL. Extensive experiments on six benchmark graph datasets, including the enormous OGB-Products graph, show that TIFA-GCL can bring a larger improvement than existing GCL methods in both transductive and inductive settings. Further experiments demonstrate the generalizability and interpretability of TIFA-GCL.

cs.LG↗

arXiv2022

On Transferability of Prompt Tuning for Natural Language Processing

Prompt tuning (PT) is a promising parameter-efficient method to utilize extremely large pre-trained language models (PLMs), which can achieve comparable performance to full-parameter fine-tuning by only tuning a few soft prompts. However, PT requires much more training time than fine-tuning. Intuitively, knowledge transfer can help to improve the efficiency. To explore whether we can improve PT via prompt transfer, we empirically investigate the transferability of soft prompts across different downstream tasks and PLMs in this work. We find that (1) in zero-shot setting, trained soft prompts can effectively transfer to similar tasks on the same PLM and also to other PLMs with a cross-model projector trained on similar tasks; (2) when used as initialization, trained soft prompts of similar tasks and projected prompts of other PLMs can significantly accelerate training and also improve the performance of PT. Moreover, to explore what decides prompt transferability, we investigate various transferability indicators and find that the overlapping rate of activated neurons strongly reflects the transferability, which suggests how the prompts stimulate PLMs is essential. Our findings show that prompt transfer is promising for improving PT, and further research shall focus more on prompts' stimulation to PLMs. The source code can be obtained from https://github.com/thunlp/Prompt-Transferability.

cs.CL↗

arXiv2022

Knowledge Inheritance for Pre-trained Language Models

Recent explorations of large-scale pre-trained language models (PLMs) have revealed the power of PLMs with huge amounts of parameters, setting off a wave of training ever-larger PLMs. However, it requires tremendous computational resources to train a large-scale PLM, which may be practically unaffordable. In addition, existing large-scale PLMs are mainly trained from scratch individually, ignoring that many well-trained PLMs are available. To this end, we explore the question how could existing PLMs benefit training large-scale PLMs in future. Specifically, we introduce a pre-training framework named "knowledge inheritance" (KI) and explore how could knowledge distillation serve as auxiliary supervision during pre-training to efficiently learn larger PLMs. Experimental results demonstrate the superiority of KI in training efficiency. We also conduct empirical analyses to explore the effects of teacher PLMs' pre-training settings, including model architecture, pre-training data, etc. Finally, we show that KI could be applied to domain adaptation and knowledge transfer.

cs.CL↗

arXiv2022

A puzzling insensitivity of magnon spin diffusion to the presence of 180$^\circ$ domain walls in a ferrimagnetic insulator

We present room-temperature measurements of magnon spin diffusion in epitaxial ferrimagnetic insulator MgAl$_{0.5}$Fe$_{1.5}$O$_{4}$ (MAFO) thin films near zero applied magnetic field where the sample forms a multi-domain state. Due to a weak uniaxial magnetic anisotropy, the domains are separated primarily by 180$^\circ$ domain walls. We find, surprisingly, that the presence of the domain walls has very little effect on the spin diffusion -- nonlocal spin transport signals in the multi-domain state retain at least 95% of the maximum signal strength measured for the spatially-uniform magnetic state, over distances at least five times the typical domain size. This result is in conflict with simple models of interactions between magnons and static domain walls, which predict that the spin polarization carried by the magnons reverses upon passage through a 180$^\circ$ domain wall.

cond-mat.mtrl-sci↗

arXiv2022

Interferometry based on quantum Kibble-Zurek mechanism

We propose an interferometry within the framework of quantum Kibble-Zurek mechanism by exemplifying two prototypical quench protocols, namely the round-trip and quarter-turn ones, on the transverse Ising and quantum $XY$ chains. Each protocol contains two linear ramps that drive the system across quantum critical point twice. The two linear ramps arouse two respective nonadiabatic critical dynamics that are well described by the quantum Kibble-Zurek mechanism. However, in combination, the two critical dynamics can interfere with each other deeply. As an effect of the interference, the dynamical phase is exposed in the final excitation probability, which leads to a quantum coherent many-body oscillation in the density of defects with predictable characteristic period. Thus such an interference is available for direct experimental observations. In the quantum $XY$ model, we show that an interference can also arise from the interplay between two different critical dynamics derived from a critical point and a tricritical point. Furthermore, we demonstrate that the interference influences the dephasing of the excited quasiparticle modes intricately by disclosing a phenomenon of multiple length scales, diagonal and off-diagonal ones, in the defect-defect correlators. It turns out that the dephased result relies on how the diagonal and off-diagonal lengths are modulated by the controllable parameter in a quench protocol.

quant-ph↗

arXiv2022

Packed Levitated Marker for Entity and Relation Extraction

Recent entity and relation extraction works focus on investigating how to obtain a better span representation from the pre-trained encoder. However, a major limitation of existing works is that they ignore the interrelation between spans (pairs). In this work, we propose a novel span representation approach, named Packed Levitated Markers (PL-Marker), to consider the interrelation between the spans (pairs) by strategically packing the markers in the encoder. In particular, we propose a neighborhood-oriented packing strategy, which considers the neighbor spans integrally to better model the entity boundary information. Furthermore, for those more complicated span pair classification tasks, we design a subject-oriented packing strategy, which packs each subject and all its objects to model the interrelation between the same-subject span pairs. The experimental results show that, with the enhanced marker feature, our model advances baselines on six NER benchmarks, and obtains a 4.1%-4.3% strict relation F1 improvement with higher speed over previous state-of-the-art models on ACE04 and ACE05.

cs.CL↗

arXiv2022

A Roadmap for Big Model

With the rapid development of deep learning, training Big Models (BMs) for multiple downstream tasks becomes a popular paradigm. Researchers have achieved various outcomes in the construction of BMs and the BM application in many fields. At present, there is a lack of research work that sorts out the overall progress of BMs and guides the follow-up research. In this paper, we cover not only the BM technologies themselves but also the prerequisites for BM training and applications with BMs, dividing the BM review into four parts: Resource, Models, Key Technologies and Application. We introduce 16 specific BM-related topics in those four parts, they are Data, Knowledge, Computing System, Parallel Training System, Language Model, Vision Model, Multi-modal Model, Theory&Interpretability, Commonsense Reasoning, Reliability&Security, Governance, Evaluation, Machine Translation, Text Generation, Dialogue and Protein Research. In each topic, we summarize clearly the current studies and propose some future research directions. At the end of this paper, we conclude the further development of BMs in a more general view.

cs.LG↗

arXiv2022

MoEfication: Transformer Feed-forward Layers are Mixtures of Experts

Recent work has shown that feed-forward networks (FFNs) in pre-trained Transformers are a key component, storing various linguistic and factual knowledge. However, the computational patterns of FFNs are still unclear. In this work, we study the computational patterns of FFNs and observe that most inputs only activate a tiny ratio of neurons of FFNs. This phenomenon is similar to the sparsity of the human brain, which drives research on functional partitions of the human brain. To verify whether functional partitions also emerge in FFNs, we propose to convert a model into its MoE version with the same parameters, namely MoEfication. Specifically, MoEfication consists of two phases: (1) splitting the parameters of FFNs into multiple functional partitions as experts, and (2) building expert routers to decide which experts will be used for each input. Experimental results show that MoEfication can conditionally use 10% to 30% of FFN parameters while maintaining over 95% original performance for different models on various downstream tasks. Besides, MoEfication brings two advantages: (1) it significantly reduces the FLOPS of inference, i.e., 2x speedup with 25% of FFN parameters, and (2) it provides a fine-grained perspective to study the inner mechanism of FFNs. The source code of this paper can be obtained from https://github.com/thunlp/MoEfication.

cs.CL↗