Jian Wu

Multi-label active learning with label correlation for image classification

2015

Label correlation analysis is very important for multi-label classification. And there is no study to measure the label correlation for example-label based active learning. In this paper, from a statistical point of view, we proposed …

Research on the Course Construction and Teaching Mode of Internet of Things System Integration

2017

Nowadays, Internet of Things (IOT) is used more and more widely, and exhibited a sustained hot situation. IOT is China's strategic emerging industries, involving computer control, wireless communications, optical communications, network, software, information security and …

An interaction consensus in group decision making under distributed trust information

2017 · DMU Open Research Archive (De Montfort University)

A theoretical interaction consensus model in group decision making with distributed linguistic trust information is proposed. To do that, the concept of distributed linguists trust function (DLTF) is defined, and then the associated operational laws …

Improving automatic source code summarization via deep reinforcement learning

2018

Code summarization provides a high level natural language description of the function performed by code, as it can benefit the software maintenance, code categorization and retrieval. To the best of our knowledge, most state-of-the-art approaches …

ResNeXt and Res2Net Structures for Speaker Verification

2020 · arXiv (Cornell University)

The ResNet-based architecture has been widely adopted to extract speaker embeddings for text-independent speaker verification systems. By introducing the residual connections to the CNN and standardizing the residual blocks, the ResNet structure is capable of …

Design and Implementation of Auxiliary System for Senior High School Physics Course

2020

培养学生的物理学核心素养,促进学生思维的形成是高中物理教学的重要目标.丰富课堂教学手段,引入计算机辅助教学提高教学质量,是近年来教学改革的重要方向.该文针对高中物理中的难点模块,采用科学计算软件Matlab设计开发了高中物理教学辅助系统,该系统具有交互性和能实现动画效果.该系统可在课堂教学中帮助学生更加直观地观察到物理过程,提高教学的效果.

Sequence-Level Confidence Classifier for ASR Utterance Accuracy and Application to Acoustic Models

2021

Scores from traditional confidence classifiers (CCs) in automatic speech recognition (ASR) systems lack universal interpretation and vary with updates to the underlying confidence or acoustic models (AMs).In this work, we build interpretable confidence scores with …

Small Models are LLM Knowledge Triggers on Medical Tabular Prediction

2024 · arXiv (Cornell University)

Recent development in large language models (LLMs) has demonstrated impressive domain proficiency on unstructured textual or multi-modal tasks. However, despite with intrinsic world knowledge, their application on structured tabular data prediction still lags behind, primarily …

ICMC-ASR: The ICASSP 2024 In-Car Multi-Channel Automatic Speech Recognition Challenge

2024

To promote speech processing and recognition research in driving scenarios, we build on the success of the Intelligent Cockpit Speech Recognition Challenge (ICSRC) held at ISCSLP 2022 and launch the ICASSP 2024 In-Car Multi-Channel Automatic …

Sequential Recommender System based on Hierarchical Attention Networks

2018

With a large amount of user activity data accumulated, it is crucial to exploit user sequential behavior for sequential recommendations. Conventionally, user general taste and recent demand are combined to promote recommendation performances. However, existing …

BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer

2019 · arXiv (Cornell University)

Modeling users' dynamic and evolving preferences from their historical behaviors is challenging and crucial for recommendation systems. Previous methods employ sequential neural networks (e.g., Recurrent Neural Network) to encode users' historical interactions from left to …

Personalized re-ranking for recommendation

2019

Ranking is a core task in recommender systems, which aims at providing an ordered list of items to users. Typically, a ranking function is learned from the labeled dataset to optimize the global performance, which …

BERT4Rec

2019

Modeling users' dynamic preferences from their historical behaviors is challenging and crucial for recommendation systems. Previous methods employ sequential neural networks to encode users' historical interactions from left to right into hidden representations for making …