Searcharxiv⌕ Search

arXiv subjects

Xuan Li

Publications and source records attributed to Xuan Li.

At least 91 records · Page 5Linked to original sources
arXiv2023

On the Evaluation of Neural Code Translation: Taxonomy and Benchmark

In recent years, neural code translation has gained increasing attention. While most of the research focuses on improving model architectures and training processes, we notice that the evaluation process and benchmark for code translation models are severely limited: they primarily treat source code as natural languages and provide a holistic accuracy score while disregarding the full spectrum of model capabilities across different translation types and complexity. In this paper, we present a comprehensive investigation of four state-of-the-art models and analyze in-depth the advantages and limitations of three existing benchmarks. Based on the empirical results, we develop a taxonomy that categorizes code translation tasks into four primary types according to their complexity and knowledge dependence: token level (type 1), syntactic level (type 2), library level (type 3), and algorithm level (type 4). We then conduct a thorough analysis of how existing approaches perform across these four categories. Our findings indicate that while state-of-the-art code translation models excel in type-1 and type-2 translations, they struggle with knowledge-dependent ones such as type-3 and type-4. Existing benchmarks are biased towards trivial translations, such as keyword mapping. To overcome these limitations, we construct G-TransEval, a new benchmark by manually curating type-3 and type-4 translation pairs and unit test cases. Results on our new benchmark suggest that G-TransEval can exhibit more comprehensive and finer-grained capability of code translation models and thus provide a more rigorous evaluation. Our studies also provide more insightful findings and suggestions for future research, such as building type-3 and type-4 training data and ensembling multiple pretraining approaches.

cs.SE↗
arXiv2023

Teacher-Student Architecture for Knowledge Distillation: A Survey

Although Deep neural networks (DNNs) have shown a strong capacity to solve large-scale problems in many areas, such DNNs are hard to be deployed in real-world systems due to their voluminous parameters. To tackle this issue, Teacher-Student architectures were proposed, where simple student networks with a few parameters can achieve comparable performance to deep teacher networks with many parameters. Recently, Teacher-Student architectures have been effectively and widely embraced on various knowledge distillation (KD) objectives, including knowledge compression, knowledge expansion, knowledge adaptation, and knowledge enhancement. With the help of Teacher-Student architectures, current studies are able to achieve multiple distillation objectives through lightweight and generalized student networks. Different from existing KD surveys that primarily focus on knowledge compression, this survey first explores Teacher-Student architectures across multiple distillation objectives. This survey presents an introduction to various knowledge representations and their corresponding optimization objectives. Additionally, we provide a systematic overview of Teacher-Student architectures with representative learning algorithms and effective distillation schemes. This survey also summarizes recent applications of Teacher-Student architectures across multiple purposes, including classification, recognition, generation, ranking, and regression. Lastly, potential research directions in KD are investigated, focusing on architecture design, knowledge quality, and theoretical studies of regression-based learning, respectively. Through this comprehensive survey, industry practitioners and the academic community can gain valuable insights and guidelines for effectively designing, learning, and applying Teacher-Student architectures on various distillation objectives.

cs.LG↗
arXiv2023

Simulating the rheology of dense suspensions using pairwise formulation of contact, lubrication and Brownian forces

Dense suspensions of solid particles in viscous liquid are ubiquitous in both industry and nature, and there is a clear need for efficient numerical routines to simulate their rheology and microstructure. Particles of micron size present a particular challenge: at low shear rates colloidal interactions control their dynamics while at high rates granular-like contacts dominate. While there are established particle-based simulation schemes for large-scale non-Brownian suspensions using only pairwise lubrication and contact forces, common schemes for colloidal suspensions generally are more computationally costly and thus restricted to relatively small system sizes. Here we present a minimal particle-based numerical model for dense colloidal suspensions which incorporates Brownian forces in pairwise form alongside contact and lubrication forces. We show that this scheme reproduces key features of dense suspension rheology near the collodial-to-granular transition, including both shear-thinning due to entropic forces at low rates and shear thickening at high rate due to contact formation. This scheme is implemented in LAMMPS, a widely-used open source code for parallelized particle-based simulations, with a runtime that scales linearly with the number of particles making it amenable for large-scale simulations.

cond-mat.soft↗
arXiv2023

On Strengthening and Defending Graph Reconstruction Attack with Markov Chain Approximation

Although powerful graph neural networks (GNNs) have boosted numerous real-world applications, the potential privacy risk is still underexplored. To close this gap, we perform the first comprehensive study of graph reconstruction attack that aims to reconstruct the adjacency of nodes. We show that a range of factors in GNNs can lead to the surprising leakage of private links. Especially by taking GNNs as a Markov chain and attacking GNNs via a flexible chain approximation, we systematically explore the underneath principles of graph reconstruction attack, and propose two information theory-guided mechanisms: (1) the chain-based attack method with adaptive designs for extracting more private information; (2) the chain-based defense method that sharply reduces the attack fidelity with moderate accuracy loss. Such two objectives disclose a critical belief that to recover better in attack, you must extract more multi-aspect knowledge from the trained GNN; while to learn safer for defense, you must forget more link-sensitive information in training GNNs. Empirically, we achieve state-of-the-art results on six datasets and three common GNNs. The code is publicly available at: https://github.com/tmlr-group/MC-GRA.

cs.LG↗
arXiv2023

Exploration of hadronization through heavy flavor production at the future Electron-Ion Collider

The future Electron-Ion Collider (EIC), which is expected to start construction at Brookhaven National Laboratory in 2025, will utilize high-luminosity high-energy electron+proton and electron+nucleus collisions to explore several fundamental questions in the high energy and nuclear physics fields. Exploring how matter is formed from quarks and gluons, which is referred to as the hadronization process, is one of the EIC science objectives. The EIC project detector design led by the ePIC collaboration can realize a series of high precision heavy flavor hadron and jet measurements. Heavy flavor jet substructure and heavy flavor hadrons inside jets, which can provide direct information about the heavy quark hadronization process, have been studied in simulation for electron+proton and electron+nucleus collisions at EIC. The associated physics projections and comparison with latest theoretical calculations will be presented.

hep-ph↗
arXiv2023

Confidence-Guided Unsupervised Domain Adaptation for Cerebellum Segmentation

The lack of a comprehensive high-resolution atlas of the cerebellum has hampered studies of cerebellar involvement in normal brain function and disease. A good representation of the tightly foliated aspect of the cerebellar cortex is difficult to achieve because of the highly convoluted surface and the time it would take for manual delineation. The quality of manual segmentation is influenced by human expert judgment, and automatic labelling is constrained by the limited robustness of existing segmentation algorithms. The 20umisotropic BigBrain dataset provides an unprecedented high resolution framework for semantic segmentation compared to the 1000um(1mm) resolution afforded by magnetic resonance imaging. To dispense with the manual annotation requirement, we propose to train a model to adaptively transfer the annotation from the cerebellum on the Allen Brain Human Brain Atlas to the BigBrain in an unsupervised manner, taking into account the different staining and spacing between sections. The distinct visual discrepancy between the Allen Brain and BigBrain prevents existing approaches to provide meaningful segmentation masks, and artifacts caused by sectioning and histological slice preparation in the BigBrain data pose an extra challenge. To address these problems, we propose a two-stage framework where we first transfer the Allen Brain cerebellum to a space sharing visual similarity with the BigBrain. We then introduce a self-training strategy with a confidence map to guide the model learning from the noisy pseudo labels iteratively. Qualitative results validate the effectiveness of our approach, and quantitative experiments reveal that our method can achieve over 2.6% loss reduction compared with other approaches.

eess.IV↗
arXiv2023

Accurate determination of low-energy eigenspectra with multi-target matrix product states

Determining the low-energy eigenspectra of quantum many-body systems is a long-standing challenge in physics. In this work, we solve this problem by introducing two novel algorithms to determine low-energy eigenstates based on a compact matrix product state (MPS) representation of the multiple targeted eigenstates. The first algorithm utilizes a canonicalization approach that takes advantage of the imaginary-time evolution of multi-target MPS, offering faster convergence and ease of implementation. The second algorithm employs a variational approach that optimizes local tensors on the Grassmann manifold, capable of achieving higher accuracy. These algorithms can be used independently or combined to enhance convergence speed and accuracy. We apply them to the transverse-field Ising model and demonstrate that the calculated low-energy eigenspectra agree remarkably well with the exact solution. Moreover, the eigenenergies exhibit uniform convergence in gapped phases, suggesting that the low-energy excited eigenstates have nearly the same level of accuracy as the ground state. Our results highlight the accuracy and versatility of multi-target MPS-based algorithms for determining low-energy eigenspectra and their potential applications in quantum many-body physics.

cond-mat.str-el↗
arXiv2023

Latest vertex and tracking detector developments for the future Electron-Ion Collider

The high-luminosity high-energy Electron-Ion Collider (EIC) to be built at Brookhaven National Laboratory (BNL) will provide a clean environment to study several fundamental questions in the high energy and nuclear physics fields. A high granularity and low material budget vertex and tracking detector is required to provide precise measurements of primary and displaced vertex and track reconstruction with good momentum and spatial resolutions. The reference design of the EIC project detector at the first Interaction Point (IP), which was selected from three submitted EIC detector proposals, enables a series of high precision heavy flavor measurements. Based on the EIC project detector reference design, the newly formed ePIC collaboration is developing the technical design for the EIC project detector towards its construction and operation. The reference design of the EIC vertex and tracking detector consists of the Monolithic Active Pixel Sensor (MAPS) based silicon vertex and tracking subsystem, the Micro-Pattern Gas Detector (MPGD) based gas tracking subsystem and the AC-Coupled Low Gain Avalanche Diode (AC-LGAD) based silicon outer tracker, and it has the track reconstruction capability in the pseudorapidity region of -3.5 to 3.5 with full azimuthal coverage. Further detector geometry optimization with a new magnet and technology down selection are underway by the ePIC collaboration. The latest ePIC vertex and tracking detector geometry and its performance evaluated in simulation will be presented. Details of the EIC vertex and tracking detector R$\&$D, which include the proposed detector technologies, prototype sensor characterization and detector mechanical design will be discussed as well.

physics.ins-det↗
arXiv2023

PAI at SemEval-2023 Task 2: A Universal System for Named Entity Recognition with External Entity Information

The MultiCoNER II task aims to detect complex, ambiguous, and fine-grained named entities in low-context situations and noisy scenarios like the presence of spelling mistakes and typos for multiple languages. The task poses significant challenges due to the scarcity of contextual information, the high granularity of the entities(up to 33 classes), and the interference of noisy data. To address these issues, our team {\bf PAI} proposes a universal Named Entity Recognition (NER) system that integrates external entity information to improve performance. Specifically, our system retrieves entities with properties from the knowledge base (i.e. Wikipedia) for a given text, then concatenates entity information with the input sentence and feeds it into Transformer-based models. Finally, our system wins 2 first places, 4 second places, and 1 third place out of 13 tracks. The code is publicly available at \url{https://github.com/diqiuzhuanzhuan/semeval-2023}.

cs.CL↗
arXiv2023

In-memory photonic dot-product engine with electrically programmable weight banks

Electronically reprogrammable photonic circuits based on phase-change chalcogenides present an avenue to resolve the von-Neumann bottleneck; however, implementation of such hybrid photonic-electronic processing has not achieved computational success. Here, we achieve this milestone by demonstrating an in-memory photonic-electronic dot-product engine, one that decouples electronic programming of phase-change materials (PCMs) and photonic computation. Specifically, we develop non-volatile electronically reprogrammable PCM memory cells with a record-high 4-bit weight encoding, the lowest energy consumption per unit modulation depth (1.7 nJ per dB) for Erase operation (crystallization), and a high switching contrast (158.5%) using non-resonant silicon-on-insulator waveguide microheater devices. This enables us to perform parallel multiplications for image processing with a superior contrast-to-noise ratio (greater than 87.36) that leads to an enhanced computing accuracy (standard deviation less than 0.007). An in-memory hybrid computing system is developed in hardware for convolutional processing for recognizing images from the MNIST database with inferencing accuracies of 86% and 87%.

physics.app-ph↗
arXiv2023

Directed Greybox Fuzzing with Stepwise Constraint Focusing

Dynamic data flow analysis has been widely used to guide greybox fuzzing. However, traditional dynamic data flow analysis tends to go astray in the massive path tracking and requires to process a large volume of data, resulting in low efficiency in reaching the target location. In this paper, we propose a directed greybox fuzzer based on dynamic constraint filtering and focusing (CONFF). First, all path constraints are tracked, and those with high priority are filtered as the next solution targets. Next, focusing on a single path constraint to be satisfied, we obtain its data condition and probe the mapping relationship between it and the input bytes through multi-byte mapping and single-byte mapping. Finally, various mutation strategies are utilized to solve the path constraint currently focused on, and the target location of the program is gradually approached through path selection. The CONFF fuzzer can reach a specific location faster in the target program, thus efficiently triggering the crash. We designed and implemented a prototype of the CONFF fuzzer and evaluated it with the LAVA-1 dataset and some real-world vulnerabilities. The results show that the CONFF fuzzer can reproduce crashes on the LAVA-1 dataset and most of the real-world vulnerabilities. For most vulnerabilities, the CONFF fuzzer reproduced the crashes with significantly reduced time compared to state-of-the-art fuzzers. On average, the CONFF fuzzer was 23.7x faster than the state-of-the-art code coverage-based fuzzer Angora and 27.3x faster than the classical directed greybox fuzzer AFLGo.

cs.CR↗
arXiv2023

PAC-NeRF: Physics Augmented Continuum Neural Radiance Fields for Geometry-Agnostic System Identification

Existing approaches to system identification (estimating the physical parameters of an object) from videos assume known object geometries. This precludes their applicability in a vast majority of scenes where object geometries are complex or unknown. In this work, we aim to identify parameters characterizing a physical system from a set of multi-view videos without any assumption on object geometry or topology. To this end, we propose "Physics Augmented Continuum Neural Radiance Fields" (PAC-NeRF), to estimate both the unknown geometry and physical parameters of highly dynamic objects from multi-view videos. We design PAC-NeRF to only ever produce physically plausible states by enforcing the neural radiance field to follow the conservation laws of continuum mechanics. For this, we design a hybrid Eulerian-Lagrangian representation of the neural radiance field, i.e., we use the Eulerian grid representation for NeRF density and color fields, while advecting the neural radiance fields via Lagrangian particles. This hybrid Eulerian-Lagrangian representation seamlessly blends efficient neural rendering with the material point method (MPM) for robust differentiable physics simulation. We validate the effectiveness of our proposed framework on geometry and physical parameter estimation over a vast range of materials, including elastic bodies, plasticine, sand, Newtonian and non-Newtonian fluids, and demonstrate significant performance gain on most tasks.

cs.CV↗
arXiv2023

Silicon detector R$\&$D for the future Electron-Ion Collider

The high-luminosity high-energy Electron-Ion Collider (EIC) to be built at Brookhaven National Laboratory (BNL) will provide a clean environment to study several fundamental questions in the high energy and nuclear physics fields. A high granularity and low material budget silicon vertex and tracking detector is required to provide precise measurements of primary and displaced vertex and track reconstruction with good momentum and spatial resolutions. The reference design of the EIC silicon vertex and tracking detector includes the Monolithic Active Pixel Sensor (MAPS) based vertex and tracking subsystem and the AC-Coupled Low Gain Avalanche Diode (AC-LGAD) based outer tracker, and it has the track reconstruction capability in the pseudorapidity region of -3.5 to 3.5 with full azimuthal coverage. Further detector geometry optimization with a new magnet based on the EIC project detector reference design are being performed by the newly formed ePIC collaboration. The latest ePIC vertex and tracking detector geometry and its performance evaluated in simulation will be presented. Details of the EIC silicon vertex and tracking detector R$\&$D, which include the proposed detector technologies, prototype sensor characterization and detector mechanical design will be discussed as well.

physics.ins-det↗
arXiv2023

Microwave heating as a universal method to transform confined molecules into armchair graphene nanoribbons

Armchair graphene nanoribbons (AGNRs) with sub-nanometer width are potential materials for fabrication of novel nanodevices thanks to their moderate direct band gaps. AGNRs are usually synthesized by polymerizing precursor molecules on substrate surface. However, it is time-consuming and not suitable for large-scale production. AGNRs can also be grown by transforming precursor molecules inside single-walled carbon nanotubes via furnace annealing, but the obtained AGNRs are normally twisted. In this work, microwave heating is applied for transforming precursor molecules into AGNRs. The fast heating process allows synthesizing the AGNRs in seconds. Several different molecules were successfully transformed into AGNRs, suggesting that it is a universal method. More importantly, as demonstrated by Raman spectroscopy, aberration-corrected high-resolution transmission electron microscopy and theoretical calculations, less twisted AGNRs are synthesized by the microwave heating than the furnace annealing. Our results reveal a route for rapid production of AGNRs in large scale, which would benefit future applications in novel AGNRs-based semiconductor devices.

cond-mat.mtrl-sci↗
arXiv2022

Teacher-Student Architecture for Knowledge Learning: A Survey

Although Deep Neural Networks (DNNs) have shown a strong capacity to solve large-scale problems in many areas, such DNNs with voluminous parameters are hard to be deployed in a real-time system. To tackle this issue, Teacher-Student architectures were first utilized in knowledge distillation, where simple student networks can achieve comparable performance to deep teacher networks. Recently, Teacher-Student architectures have been effectively and widely embraced on various knowledge learning objectives, including knowledge distillation, knowledge expansion, knowledge adaption, and multi-task learning. With the help of Teacher-Student architectures, current studies are able to achieve multiple knowledge-learning objectives through lightweight and effective student networks. Different from the existing knowledge distillation surveys, this survey detailedly discusses Teacher-Student architectures with multiple knowledge learning objectives. In addition, we systematically introduce the knowledge construction and optimization process during the knowledge learning and then analyze various Teacher-Student architectures and effective learning schemes that have been leveraged to learn representative and robust knowledge. This paper also summarizes the latest applications of Teacher-Student architectures based on different purposes (i.e., classification, recognition, and generation). Finally, the potential research directions of knowledge learning are investigated on the Teacher-Student architecture design, the quality of knowledge, and the theoretical studies of regression-based learning, respectively. With this comprehensive survey, both industry practitioners and the academic community can learn insightful guidelines about Teacher-Student architectures on multiple knowledge learning objectives.

cs.LG↗
arXiv2022

Forward silicon tracking detector developments for the future Electron-Ion Collider

The future Electron-Ion Collider (EIC) will utilize a series of high-luminosity high-energy electron+proton ($e+p$) and electron+nucleus ($e+A$) collisions to explore the inner structure of nucleon and nucleus and the matter formation process. Heavy flavor hadron and jet measurements at the EIC will play an essential role in determining the nucleon/nucleus parton distribution function and heavy quark hadronization process in not well constrained kinematic regions. A high granularity and low material budget forward silicon tracker will enable precise forward heavy flavor measurements at the EIC, which have enhanced sensitivities to access these kinematic extremes. A Forward Silicon Tracker (FST) detector is under design and R$\&$D for the EIC. Two advanced silicon technologies, the Depleted Monolithic Active Pixel Sensor (DMAPS) and the AC coupled Low Gain Avalanche Diode (AC-LGAD), which can provide fine spatial and timing resolutions, have been considered as candidates for the EIC silicon tracking detector. Progresses and results about the FST conceptual design and ongoing DMAPS and LGAD detector R$\&$D will be presented. The path towards an integrated EIC detector will be discussed as well.

physics.ins-det↗
arXiv2022

Topology Optimization with Frictional Self-Contact

Contact-aware topology optimization faces challenges in robustness, accuracy, and applicability to internal structural surfaces under self-contact. This work builds on the recently proposed barrier-based Incremental Potential Contact (IPC) model and presents a new self-contact-aware topology optimization framework. A combination of SIMP, adjoint sensitivity analysis, and the IPC frictional-contact model is investigated. Numerical examples for optimizing varying objective functions under contact are presented. The resulting algorithm proposed solves topology optimization for large deformation and complex frictionally contacting scenarios with accuracy and robustness.

math.OC↗