Search

Your search keyword '"Chen, Tianyu"' showing total 71 results
Start Over Author "Chen, Tianyu" Publisher ieee
71 results on '"Chen, Tianyu"'

15. Learning Music Sequence Representation From Text Supervision

Author

Chen, Tianyu, Xie, Yuan, Zhang, Shuai, Huang, Shaohan, Zhou, Haoyi, and Li, Jianxin

Subjects
FOS: Computer and information sciences, Computer Science - Machine Learning, Sound (cs.SD), Audio and Speech Processing (eess.AS), FOS: Electrical engineering, electronic engineering, information engineering, Computer Science - Sound, Electrical Engineering and Systems Science - Audio and Speech Processing, Machine Learning (cs.LG)
Abstract

Music representation learning is notoriously difficult for its complex human-related concepts contained in the sequence of numerical signals. To excavate better MUsic SEquence Representation from labeled audio, we propose a novel text-supervision pre-training method, namely MUSER. MUSER adopts an audio-spectrum-text tri-modal contrastive learning framework, where the text input could be any form of meta-data with the help of text templates while the spectrum is derived from an audio sequence. Our experiments reveal that MUSER could be more flexibly adapted to downstream tasks compared with the current data-hungry pre-training method, and it only requires 0.056% of pre-training data to achieve the state-of-the-art performance.

Published
2022

16. A Novel Bluetooth-Odometer-Aided Smartphone-Based Vehicular Navigation in Satellite-Denied Environments.

Author

Wen, Zeyang, Yang, Gongliu, Cai, Qingzhong, and Chen, Tianyu

Subjects
SMARTPHONES, MICROELECTROMECHANICAL systems, NAVIGATION, AUTOMOTIVE navigation systems, UNITS of measurement, MOBILE apps, PROBLEM solving, ARTIFICIAL satellites in navigation
Abstract

This article investigates a novel Bluetooth odometer (BOD)-aided smartphone-based vehicular navigation system for satellite-denied environments. First, a wheel-mounted BOD is designed to measure rotational speed and send data to smartphones by Bluetooth transmission. Second, a nonlinear smartphone-based microelectromechanical system (MEMS) inertial measurement unit (IMU)/BOD error model is derived to solve the problem of azimuth divergence. Then, a position delay time model is proposed to compensate for the delay time of Bluetooth transmission. An Android app is developed to receive the BOD data from Bluetooth transmission, and the proposed BOD smartphone-based vehicular navigation algorithm provides positioning information in real time. In field tests, positioning accuracy reaches 0.38%D by using the proposed MEMS-IMU/BOD model. The designed scheme can be used as a low-cost and convenient solution for smartphone-based civilian vehicular autonomous navigation systems without changing the habits of drivers. [ABSTRACT FROM AUTHOR]

Published
2023
Full Text
View/download PDF

22. Interpretable Utility-based Models Applied to the FightingICE Platform

Author

Chen, Tianyu, Richoux, Florian, Torres, Javier, Inoue, Katsumi, Richoux, Florian, The Graduate University for Advanced Studies, japan (SOKENDAI), National Institute of Advanced Industrial Science and Technology (AIST), Japanese French Laboratory for Informatics (JFLI), National Institute of Informatics (NII)-The University of Tokyo (UTokyo)-Sorbonne Université (SU)-Centre National de la Recherche Scientifique (CNRS), Chercheur indépendant, and National Institute of Informatics (NII)

Subjects
Machine Learning, [INFO.INFO-AI] Computer Science [cs]/Artificial Intelligence [cs.AI], ComputingMilieux_PERSONALCOMPUTING, Interpretable Results, Fighting Games, Utility-based AI, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract

International audience; One task of game designers is to give NPCs fun behaviors, where "fun" can have many different manifestations. Several classic methods exist in Game AI to model NPCs' behaviors; one of them is utilitybased AI. Utility functions constitute a powerful tool to define behaviors but can be tedious and timeconsuming to make and tune correctly until the desired behavior is achieved. Here, we propose a method to learn utility functions from data collected after some human-played games, to recreate a target behavior. Utility functions are modeled using Interpretable Compositional Networks, allowing us to get interpretable results, unlike regular neural networks. We show our method can handle noisy data and learn utility functions able to credibly reproduce different target behaviors, with a median accuracy from 64.5% to 83.7%, using the FightingICE platform, an environment for AI agent competitions. We believe our method can be useful to game designers to quickly prototype NPCs' behaviors, and even to define their final utility functions.

Published
2021

29. Test Method of Connector Terminal Temperature Rise based on New Test Platform

Author

Zhou Bo, Yang Hao, Zhao Haishan, Sun Ping, Chen Tianyu, Lin Dajie, and Sun Yue

Subjects
ComputingMilieux_GENERAL, Cable gland, Safe operation, Terminal (electronics), Computer science, Test platform, Electrical connector, Key (cryptography), Test method, Automotive engineering, Test (assessment)
Abstract

The temperature rise test of rail vehicle electrical connector terminal is the key to ensure the safe operation of rail vehicle. In this paper, a new test method of connector terminal temperature rise based on new test platform is proposed on the basis of summarizing the existing test methods. The factors affecting the accuracy of temperature rise test are discussed. Experimental results verify the effectiveness of the new method.

Published
2021

37. A Lightweight Full Entropy TRNG With On-Chip Entropy Assurance.

Author

Chen, Tianyu, Ma, Yuan, Lin, Jingqiang, Cao, Yuan, Lv, Na, and Jing, Jiwu

Subjects
*MONTE Carlo method, *RANDOM number generators, *MODEL theory, *QUALITY of service, *PERFORMANCE technology
Abstract

True random number generator (TRNG) as one essential hardware primitive is widely used in cryptography, Monte Carlo simulation, and gambling. To evaluate the security of TRNG, the entropy of the TRNG’s output is usually estimated by the stochastic model in theory or measured off-chip after fabrication. However, the sufficiency of entropy is difficult to be guaranteed in practice due to the facts: 1) the inaccuracy of the model-based jitter measurement method; 2) the variations of the chip manufacturing process and operating environments (such as supply voltage and temperature); and 3) malicious attacks. In this work, we design a novel TRNG architecture with on-chip entropy assurance to properly solve practical security problems. In the design, we propose an on-chip entropy estimator for measuring independent jitter to quantify true randomness, which enables continuous monitoring of TRNG at runtime. Furthermore, with the cooperation of the proposed on-chip entropy estimator and a rational self-adaptive mechanism, the designed TRNG can steadily generate bitstreams with sufficient entropy (≥ 0.999 per bit) against PVT variations. We implement the TRNG architecture in FPGAs with different technology nodes (45 and 65 nm) and SMIC 130 nm chips. Experimental results validate that the designed TRNG has an excellent performance in terms of technology independence and environmental robustness. The generated bitstreams pass the NIST SP800-22 and Diehard statistical test suites successfully without any post-processing. [ABSTRACT FROM AUTHOR]

Published
2021
Full Text
View/download PDF

38. Review of Interference Suppression Algorithms for Passive Sonar

Author

Chen Wen, Chen Tianyu, Wu Yanqun, Hu Zhengliang, and Zhang Wen

Subjects
010505 oceanography, Computer science, 020208 electrical & electronic engineering, Array processing, 02 engineering and technology, Interference (wave propagation), 01 natural sciences, Sonar, Subspace decomposition, 0202 electrical engineering, electronic engineering, information engineering, Underwater, Joint (audio engineering), Algorithm, Subspace topology, 0105 earth and related environmental sciences
Abstract

Interference suppression (IS) algorithms have been extensively discussed for it can reveal the target signals among the interference. Numerous IS algorithms have been proposed for passive sonar. To appropriately apply them in different scenarios, a novel category is proposed in this paper to analyze them and classify them into three methods (the adaptive array processing technology based nulling method, the subspace separation based IS method and the joint time-space-frequency based IS method). In the meantime, the complexity, applicable occasions, strengths and weaknesses of three types of IS algorithms are analyzed.

Published
2019

39. Split Beam Cross Correlation for A Hybrid Acoustic Vector Sensor Array

Author

Wu Yanqun, Chen Tianyu, and Hu Zhengliang

Subjects
Vector sensor, Cross-correlation, 010505 oceanography, Computer science, Acoustic vector sensor, Acoustics, media_common.quotation_subject, 020208 electrical & electronic engineering, Energy information, 02 engineering and technology, Ambiguity, 01 natural sciences, Directivity, Broadband, 0202 electrical engineering, electronic engineering, information engineering, Beam (structure), 0105 earth and related environmental sciences, media_common
Abstract

A hybrid acoustic vector sensor array (AVSA) is consisted mainly of pressure hydrophones and a few vector sensors within. It aims at acquiring good left-right ambiguity elimination performance, while avoiding the disadvantages and complexity of full-rank AVSA which is completely made up with vector sensors. This paper proposes a split beam cross-correlation (SBC) method for the hybrid AVSA, which cross-correlates the beam outputs of the pressure-only subarray and vector sensor subarray in the hybrid AVSA to get the unambiguous direction and energy information of targets with only one vector sensor or a few. Simulation results validate the performance of the SBC method and show that the proposed method for hybrid AVSA can completely suppress the false target and eliminate the ambiguity for broadband targets at a comparatively low cost.

Published
2019

40. Global Abnormal Event Detection in Video via Motion Information Entropy

Author

Li Beichen, Chunping Hou, Li Yao, Chen Tianyu, Zhipeng Wang, and Mei Song

Subjects
Computational complexity theory, Computer science, business.industry, Gaussian, ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION, Optical flow, Pattern recognition, Information needs, 02 engineering and technology, symbols.namesake, 0202 electrical engineering, electronic engineering, information engineering, Task analysis, symbols, Detection performance, Entropy (information theory), 020201 artificial intelligence & image processing, Artificial intelligence, business
Abstract

Abnormal event detection is a rapid development task in video analysis, which is aimed to distinguish abnormal and normal events in surveillance videos. As the normal and abnormal events have some similarities, more discriminating methods or motion information need to be explored. In this paper, to detect escaping people in different scenes, a global abnormal event detection method is proposed based on motion information entropy. To be specific, the proposed method utilizes the directions and magnitudes of motion vectors extracted from optical flow field to calculate the motion information entropy which represents the uncertainty of motion information. The normal samples are distributed around the center of a Gaussian distribution while the abnormal ones are distributed on the side. Experimental result shows that the proposed method has better detection performance than most of other classic methods with much lower computational complexity.

Published
2018

45. Electromagnetic Compatibility Analysis of an Induction Motor Drive With Integrated Power Converter.

Author

Chen, Tianyu, Caicedo-Narvaez, Carlos, Wang, Haokun, Moallem, Mehdi, Fahimi, Babak, and Kiani, Morgan

Subjects
*INDUCTION machinery, *ELECTROMAGNETIC compatibility, *MAGNETIC flux leakage, *WIDE gap semiconductors, *ELECTRONIC circuits, *INDUCTION motors, *ELECTROMAGNETIC devices
Abstract

With the development of high-temperature wide bandgap power semiconductors, such as SiC MOSFET and GaN HEMT, it becomes possible to integrate the power converter directly into motors. However, the leakage magnetic flux in the motor may cause interference with the operation of the electronic circuit. In this article, the electromagnetic compatibility of the power converter integrated in an induction motor is investigated. The simulation shows that the leakage magnetic flux of the end-winding will penetrate into the nearby area, and it will cause interference with the Hall-effect current sensors if the power converter is mounted in the space between the motor end cap and the end-winding. A prototype of the GaN-HEMT-based power converter is built and integrated into a 3-hp induction motor. The leakage magnetic flux is scanned by the linear Hall-effect sensors and the measured result is aligned with the finite-element simulation. [ABSTRACT FROM AUTHOR]

Published
2020
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results