Phd Thesis

Mikael Kågebäck, (PhD thesis, Department of Computer Science and Engineering (Chalmers))

PhD seminar

Date & time: ~15-30 September. (Exact date TBA)
Location: Chalmers, Göteborg, Sweden. (Room TBA)

Opponent

TBA

Supervisiors

Main supervisor: Devdatt Dubhashi
Co-supervisors: Richard Johansson, Shalom Lappin


List of papers, including venue and contributions:

Paper I: Neural context embeddings for automatic discovery of word senses (NAACL 2015 workshop on Vector Space Modeling for NLP) [pdf]

  • Main author.
  • Developed the main idea.
  • Wrote ~50% of the text.
  • Implemented ~50% of the experiments.

Paper II: Learning to Play Guess Who? and Inventing a Grounded Language as a Consequence (NIPS workshop on deep reinforcement learning) [pdf]

  • Initiated the project.
  • Supervised the main author
  • Contributed towards the manuscript (abstract, introduction, and conclusions)
  • Contributed towards the technical contribution of the paper.

Paper III: Word Sense Disambiguation using a Bidirectional LSTM (Coling 2016 Workshop on Cognitive Aspects of the Lexicon (CogALex-V)) [pdf]

  • Developed the main idea.
  • Wrote 90% of the text.

Paper IV: Extractive Summarization using Continuous Vector Space Models (EACL 2014 workshop on continuous vector space models and their compositionality (CVSC)) [pdf]

  • Main author.
  • Developed the main idea.
  • Wrote ~80% of the text.
  • Implemented ~50% of the experiments.

Paper V: Extractive Summarization by Aggregating Multiple Similarities (RANLP 2015) [pdf]

  • Second author.
  • Multiplicative interaction between kernels.
  • Wrote ~20% of the text.
  • Implemented ~20% of the experiments (the parts relating to word embeddings).

Paper VI: Disentangled Representations for Manipulation of Sentiment in Text (NIPS workshop on Learning Disentangled Features: from Perception to Control) [pdf]

  • Formulated the research question.
  • Supervised the main authors.
  • Wrote ~50% of the text.

Paper VII: Disentangled activations in deep networks (NIPS workshop on Learning Disentangled Features: from Perception to Control and under submission to ICLR) [nips, iclr]

  • Contributed the main idea of the paper.
  • Wrote ~50% of the text.
  • Implemented two of the three experiments.

Paper VIII: Optimal Transport of deep representations (Ongoing work planned for ACL)

Paper IX: Acquiring colour terms by solving a problem (Ongoing work planned for ACL)


DEMO

Summarisation demo