Program

9th (Thu)


     09:00 - 09:10    |     Opening
                                    Key-Sun Choi
     ??:?? - ??:??    |     Talk-1: DeepPavlov: Open-Source Library for Dialogue Systems.
                                    Mikhail S. Burtsev
     ??:?? - ??:??    |     Talk-2: Deep neural network-based video summarization.
                                    Yuta Nakashima
     ??:?? - ??:??    |     Talk-3: Effective Semantics for Engineering NLP systems
                                    André Freitas                                                


Keynote Speakers

Talk-1: Mikhail S. Burtsev
Head of the of the Neural Networks and Deep Learning Lab.
Title 
    DeepPavlov: Open-Source Library for Dialogue Systems.
Abstract
    Adoption of messaging communication and voice assistants has grown rapidly in the last years. This 
creates a demand for tools that speed up prototyping of feature-rich dialogue systems. An open-source 
library DeepPavlov is tailored for development of conversational agents. The library prioritises efficiency, 
modularity, and extensibility with the goal to make it easier to develop dialogue systems from scratch 
and with limited data available. It supports modular as well as end-to-end approaches to implementation 
of conversational agents. Conversational agent consists of skills and every skill can be decomposed into 
components. Components are usually models which solve typical NLP tasks such as intent classification, 
named entity recognition or pre-trained word vectors. Sequence-to-sequence chit-chat skill, question 
answering skill or task-oriented skill can be assembled from components provided in the library.

Talk-2: Yuta Nakashima
Associate Professor of Osaka University.
Title 
    Deep neural network-based video summarization
Abstract
     Video summarization is a technique to compactly represent a long and redundant video, and various 
methods have been proposed so far. In recent years, a new approach has emerged that uses deep 
neural networks to “understand” a given video. In this talk, I will overview the recent approach and 
introduce our video summarization method that maps a sentence and a video into the same space 
for extracting higher-level semantics of the video. I also would like to mention about the difficulty of 
handling sequence data (e.g., videos and text) with deep neural networks.

Talk-3: André Freitas
Lecturer of the University of Manchester.
Title 
    Effective Semantics for Engineering NLP Systems
Abstract
     At the center of many Natural Language Processing (NLP) applications is the requirement for capturing 
and interpreting commonsense and domain-specific knowledge at scale. The selection of the right semantic 
and knowledge representation model plays a strategic role for building NLP systems ( e.g. Question Answering, 
Sentiment Analysis, Semantic Search) which effectively work with real data.
    In this talk, we will provide an overview of emerging trends in semantic representation for building NLP 
systems which can cope with large-scale and heterogeneous textual data. Based on empirical evidence, we 
will provide a description of the strengths and weaknesses of different representation perspectives, aiming 
towards a synthesis: ‘a semantic model to rule them all’.