-
03
2020-01
2020年系列学术活动(第三场)——北京交通大学于剑教授学术报告
报告题目:基于认知的机器学习公理化报告内容:在大数据时代,因应用需求的驱动,大量新机器学习方法不断产生。 这些新算法理论依据各异,彼此之间的关系极其复杂,对学习算法的使用者要求极高。但是, 儿童的学习能力虽高, 却不能掌握现今机器学习的理论。 是否能够提出一套符合人类认知的机器学习理论,是当前一个亟待解决的问题。本次报告试图提出一个统一基于认知的机器学习公理化框架, 其基本假设是: 归哪类,像哪类...
-
03
2020-01
2020年系列学术活动(第二场)——澳大利亚蒙纳士大学刘铭博士学术报告
报告题目:Learning Deep Neural Networks for Hard Natural Language Problems报告摘要:Deep learning has revolutionized the way that many tasks are tackled in Natural Language Processing. The talk will cover some recent advances in deep learning for NLP, including embeddings, encoder-decoder architecture, attentions and language modelling. Then, I will introduce some of the research work in Mo...
-
30
2019-12
2020年系列学术活动(第一场)——新加坡管理大学Amy研究员学术报告
One-Class Order Embedding Learning for Dependency Relation Prediction Abstract: Most of today's representation learning techniques aim to learn entity representations that are semantics-preserving, i.e., semantically similar entities are mapped into a nearby area in the semantic embedding space. In this study, we aim to learn order-preserving representations for entities, i.e., antisymmetri...
-
26
2019-12
澳大利亚国立大学杜岚博士学术报告
报告题目:Multi-label Learning with/without Zero-shot Classes报告内容:Multi-label learning refers to the problem of learning to assign a subset of relevant labels to each object, drawn from a large set of candidate labels. Each object is thus associated with a binary label vector, which denotes the presence/absence of each of the candidate labels. In real-world application, such as image/do...
-
21
2019-12
美国哥伦比亚大学陈博源博士学术报告
报告题目:Machine Theory of Behavior and Mind 报告摘要:We have seen impressive results from Machine Learning that enable machines to recognize objects very accurately, translate between multiple languages, and manipulate various objects with high success rate. However, there is still a gap to build machines that can operate in unconstrained environment. Humans are able to understand the goa...
-
20
2019-12
墨尔本大学马兴军助理讲师学术报告
报告题目:Adversarial Machine Learning: an intruduction and tutorial 报告摘要:Deep learning has become increasingly popular in the past few years. This is largely attributed to a family of powerful models called deep neural networks (DNNs). With many stacked layers, and millions of neurons, DNNs are capable of learning complex non-linear mappings, and have demonstrated near or even surpassi...
-
20
2019-12
南洋理工大学杨杰龙博士学术报告
报告题目:On Truth Finding in Multi-agent Networks 报告摘要:A problem named “truth finding” will be discussed in this talk. In this problem, an event is a phenomenon of interest and our goal is to infer the states of events (e.g., articles, movies, and signals) through biased observations of multiple agents (e.g., sensors and individuals) in a network. The local bias of each agent is unkn...
-
16
2019-12
中国计算机学会上海分部主席白硕教授学术报告
报告题目:Towards Precise Parsing: What and How 报告摘要:自然语言的精准解析是人工智能皇冠上的明珠,也是很多实际应用领域的共同期盼。本 报告结合知识图谱中事件知识处理研究和应用落地的共同需求,介绍报告者在自然语言精准解析领域几十年研究成果——关联语法(what)及其分析技术(how),讨论精准解析与事件知识处理的衔接,并探索这些技术在金融领域的应用前景。 报告人介绍:白硕,金融科技专家,中国计算机...
-
10
2019-12
密歇根州立大学李照剑教授学术报告
ZHAOJIAN LIAssistant ProfessorDepartment of Mechanical EngineeringMichigan State University Talk IFuture Mobility: Integrating Data-Driven and Control Methods into Automotive Decision-Making Systems2019年12月16日 星期一9:00 a.m -11:00 a.m 南岭校区先进控制基地A区2楼报告厅ABSTRACT: Data is everywhere. Modern vehicles are equipped with hundreds of sophisticated sensors that offer necessary inf...
-
26
2019-11
澳大利亚莫纳什大学赵贺博士学术报告
报告题目:基于元数据主题模型的短文本分析和主题结构挖掘报告摘要:作为一种概率生成模型,主题模型主要应用于离散数据并假设输入数据由若干“潜在因素”生成。在文本分析中,这些“潜在因素”通常包含一些特定的含义,而每一种含义又可以被一组特定的词所解释,所以这些“潜在因素”又可以被称为“主题”。因此,一篇文章里的词频可以视作是由若干不同含义的主题共同生成的,而每种主题在文章中所占的比重也不同。在近二十...