@harm_devries

PhD candidate @ UMontreal.

 Financial District, Manhattan
Joined December 2010
  • Tweets18
  • Following48
  • Followers328
Visually Grounded Interaction and Language: accepted works together with spotlights are available: nips2017vigil.github.io at #nips2017 #nips

Visually-Grounded Interaction and Language (ViGIL)

NIPS 2017 Workshop, Long Beach, California, USA

nips2017vigil.github.io
12
3
24
Harm de Vries retweeted
Abhishek Das
@abhshkdz30 Oct 2017
#NIPS2017 Vis-Grounded Interaction & Lang wrkshp deadline Nov 17! Consider submitting your #ICLR2018 / #CVPR2018 drafts + other recent wrk
1
2
2
10
Extended version of our paper on tackling visual reasoning with conditional linear transformations is out: arxiv.org/abs/1709.07871

FiLM: Visual Reasoning with a General Conditioning Layer

We introduce a general-purpose conditioning method for neural networks called FiLM: Feature-wise Linear Modulation. FiLM layers influence neural network computation via a simple, feature-wise...

arxiv.org
4
19
Code: github.com/ethanjperez/film. Extended analysis includes examining the compositional nature of learned question embeddings.

GitHub - ethanjperez/film: FiLM: Visual Reasoning with a General Conditioning Layer

FiLM: Visual Reasoning with a General Conditioning Layer - ethanjperez/film

github.com
1
13
Please consider submitting your work to our NIPS17 workshop on Visually-Grounded Interaction and Language! nips2017vigil.github.io xitter.cf/abhshkdz/status/911695…
Abhishek Das
@abhshkdz23 Sep 2017
Excited to be co-orgnizing the NIPS 2017 wrkshop on Visually-Grounded Interaction & Language! Submissions due Nov 3! nips2017vigil.github.io

Visually-Grounded Interaction and Language (ViGIL)

NIPS 2017 Workshop, Long Beach, California, USA

nips2017vigil.github.io
16
32
New paper with @univ_lille & @UMontreal uses RL to learn grounded natural language via image-based game 'Guess What' arxiv.org/abs/1703.05423

End-to-end optimization of goal-driven and visually grounded...

End-to-end design of dialogue systems has recently become a popular research topic thanks to powerful tools such as encoder-decoder architectures for sequence-to-sequence learning. Yet, most...

arxiv.org
66
3
160
CVPR paper on dialogue with images accepted! Congrats to @harm_devries, @apsarathchandar, @AaronCourville et al. guesswhat.ai/
17
75
Harm de Vries retweeted
Nando de Freitas
@NandoDF2 Jan 2017
Rise of the machines can make life better. The Times, excellent article by Emma Duncan on universal income @1843mag thetimes.co.uk/article/rise-…
1
27
3
50
Harm de Vries retweeted
Eric Jang
@ericjang1121 Dec 2016
Nice article by @laurent_dinh on posterior inference for samples from Gumbel-Max trick laurent-dinh.github.io/2016/…
15
50
Harm de Vries retweeted
Sander Dieleman
@sedielem9 Dec 2016
Best company party at #nips2016: definitely #rocketai
7
24
Harm de Vries retweeted
David Pfau@pfau9 Dec 2016
Proud to have been part of the #RocketAI launch tonight. Sorry all the seed funding will be going to legal fees.
9
1
51
Harm de Vries retweeted
Andrej Karpathy
@karpathy9 Dec 2016
Best party of #nips2016 award goes to #rocketai (rocketai.org/). Definitely a company to watch closely.
2
41
2
124
"GuessWhat?! Visual object discovery through multi-modal dialogue," de Vries et al.: arxiv.org/abs/1611.08481

GuessWhat?! Visual object discovery through multi-modal dialogue

We introduce GuessWhat?!, a two-player guessing game as a testbed for research on the interplay of computer vision and dialogue systems. The goal of the game is to locate an unknown object in a...

arxiv.org
19
1
32
Load more