Overview

The 2nd Workshop on Representation Learning for NLP (RepL4NLP) will be held on August 3, and hosted by the 55th Annual Meeting of the Association for Computational Linguistics (ACL) in Vancouver, Canada. The workshop is being organised by Phil Blunsom, Antoine Bordes, Kyunghyun Cho, Shay Cohen, Chris Dyer, Edward Grefenstette, Karl Moritz Hermann, Laura Rimell, Jason Weston, and Scott Wen-tau Yih, and will be sponsored by DeepMind, Microsoft Research, and Facebook AI Research.

The 2nd Workshop on Representation Learning for NLP aims to continue the success of the 1st Workshop on Representation Learning for NLP (about 50 submissions and over 250 attendees; second most attended collocated event at ACL'16 after WMT) which was introduced as a synthesis of several years of independent *CL workshops focusing on vector space models of meaning, compositionality, and the application of deep neural networks and spectral methods to NLP. It provides a forum for discussing recent advances on these topics, as well as future research directions in linguistically motivated vector-based models in NLP.

Key Dates
  • Deadline for paper submission: April 28, 2017 Extended deadline
  • Notification of acceptance: May 23, 2017
  • Camera ready submission due: May 30, 2017
  • Early registration deadlineJune 23, 2017
  • Workshop: 3 August 2017
Keynote Speakers
  • Jacob Eisenstein, Georgia Tech Jacob Eisenstein is an Assistant Professor in the School of Interactive Computing at Georgia Tech, where he leads the Computational Linguistics Laboratory. He works on machine learning approaches to understanding human language. He is especially interested in non-standard language, discourse, computational social science, and statistical machine learning.
    Keynote: Learning Representations of Social Meaning Language plays a critical role in structuring human relationships, while marking social properties of the speaker/writer, audience, and communicative situation. With the increasing availability of big social media datasets, computational linguists have begun to join with sociolinguists in working to elucidate language's social dimension. However, this promising synthesis is threatened by a theoretical mismatch between these two disciplines. Much of the research in the emerging field of computational sociolinguistics involves social-theoretical models that uncritically assign individuals to broad categories such as man/woman, black/white, northern/southern, and urban/rural. Meanwhile, sociolinguists have worked for decades to elaborate a more nuanced view of identity and social meaning, but it has proven difficult to reconcile these rich theoretical models with scalable quantitative research methods. In this talk, I will ask whether representation learning can help to bridge this gap. The key idea is to use learned representations to mediate between linguistic data and socially relevant metadata. I will describe applications of this basic approach in the context of clustering, latent variable models, and neural networks, with applications to gender, multi-community studies, and social network analysis.
  • Sanja Fidler, Toronto Sanja Fidler is an Assistant Professor at the Department of Computer Science, University of Toronto. Previously she was a Research Assistant Professor at TTI-Chicago, a philanthropically endowed academic institute located in the campus of the University of Chicago. She completed her PhD in computer science at University of Ljubljana in 2010, and was a postdoctoral fellow at University of Toronto during 2011-2012. In 2010 she visited UC Berkeley. She has served as a Program Chair of the 3DV conference, and as an Area Chair of CVPR, ICCV, EMNLP, ICLR, and NIPS. She received the NVIDIA Pioneer of AI award. Her main research interests are object detection, 3D scene understanding, and the intersection of language and vision. Keynote: Learning Joint Embeddings of Vision and Language A successful autonomous system needs to not only understand the visual world but also communicate its understanding with humans. To make this possible, language can serve as a natural link between high level semantic concepts and low level visual perception. In this talk, I'll discuss recent work in the domain of vision and language, covering topics such as image/video captioning and retrieval, and question-answering. I’ll also talk about our recent work on task execution via language instructions.
  • Mirella Lapata, Edinburgh Mirella Lapata is a Professor at the School of Informatics at the University of Edinburgh. Her recent research interests are in natural language processing. She serves as an associate editor of the Journal of Artificial Intelligence Research (JAIR). She is the first recipient (2009) of the British Computer Society and Information Retrieval Specialist Group (BCS/IRSG) Karen Sparck Jones award. She has also received best paper awards in leading NLP conferences and financial support from the EPSRC (the UK Engineering and Physical Sciences Research Council) and ERC (the European Research Council). Keynote: "A million ways to say I love you" or Learning to paraphrase with Neural Machine Translation Recognizing and generating paraphrases is an important component in many natural language processing applications. A well-established technique for automatically extracting paraphrases leverages bilingual corpora to find meaning-equivalent phrases in a single language by "pivoting" over a shared translation in another language. In the first part of the talk I will revisit bilingual pivoting in the context of neural machine translation and present a paraphrasing model based purely on neural networks. The proposed model represents paraphrases in a continuous space, estimates the degree of semantic relatedness between text segments of arbitrary length, and generates paraphrase candidates for any source input. In the second part of the talk I will illustrate how neural paraphrases can be seamlessly integrated in models of question answering and summarization, achieving competitive results across datasets and languages.Recognizing and generating paraphrases is an important component in many natural language processing applications. A well-established technique for automatically extracting paraphrases leverages bilingual corpora to find meaning-equivalent phrases in a single language by "pivoting" over a shared translation in another language. In the first part of the talk I will revisit bilingual pivoting in the context of neural machine translation and present a paraphrasing model based purely on neural networks. The proposed model represents paraphrases in a continuous space, estimates the degree of semantic relatedness between text segments of arbitrary length, and generates paraphrase candidates for any source input. In the second part of the talk I will illustrate how neural paraphrases can be seamlessly integrated in models of question answering and summarization, achieving competitive results across datasets and languages.
  • Alona Fyshe, University of Victoria 
    Alona Fyshe is an Assistant Professor in the Computer Science Department at the University of Victoria and a CIFAR Global Scholar. Alona received her BSc and MSc in Computing Science from the University of Alberta, and a PhD in Machine Learning from Carnegie Mellon University. Alona uses machine learning to leverage large amounts of text and neuroimaging data to understand how people mentally combine words to create higher-order meaning. Alona is also interested in improving machine learning models by understanding the brain’s structure and function. Keynote: Representations in the Brain What can the brain tell us about computationally-learned representations of words, phrases and beyond? And what can those computational representations tell us about the brain? In this talk I will describe several brain imaging experiments that explore the representation of language meaning in the brain, and relate those brain representations to computationally learned representations language meaning.
Topics
  • Distributional compositional semantics
  • Analysis of language using eigenvalue, singular value and tensor decompositions
  • Latent-variable and representation learning for language
  • Neural networks and deep learning in NLP
  • Word embeddings and their applications
  • Spectral learning and the method of moments in NLP
  • Language modeling for automatic speech recognition, statistical machine translation, and information retrieval
  • The role of syntax in compositional models
  • Language modeling for logical and natural reasoning
  • Integration of distributional representations with other models
  • Multi-modal learning for distributional representations
  • Knowledge base embedding