Roelof Pieters

More info on my work with DeepDreaming / Visualizing the internal representations of Deep Neural Nets also here

(Disclaimer: This is a personal website. All ramblings are my own :) )

General Information

I'm a PhD student at the Computer Science and Communication department at KTH, as well as working for an exciting data-science consultancy: Graph Technologies


My major fields of interest are Deep Learning, Natural Language Processing, Distributed Semantics, Data Visualization, Graphs, Machine Translation, Recommender and Search Systems.
My supervisors are Viggo Kann and Jussi Karlgren.

Meetups

• Stockholm Natural Language Processing Meetup
• Stockholm Deep Learning Meetup

Talks/Lectures

Reading Groups

• Usually Fridays at 13.30-15.00, room 304 | Vision/Language Deep Learning reading group

2016

• May 2 at KTH | Guest Lecture #2 DD2427 | Multi-Modal Embeddings: from Discriminative to Generative Models and Creative AI | slides
• May 2 at KTH | Guest Lecture #1 DD2427 | Deep learning for natural language embeddings | slides
• April 26 at KTH | Guest Lecture DD2476 | Multi-modal Retrieval and Generation with Deep Distributed Models | slides

2015

• December 3 at KTH | Guest Lecture DD2418 | Deep Learning for Natural Language Processing: Word Embeddings | slides
• December 1 | Imperial College London, BICV map | Creative AI & Multimodality: looking ahead | slides
• September 23 | Sthlm Deep Learning Meetup | Deep Learning & Creativity (Language and Vision) | abstract | video
• June 21 | PyData London | Lightning Talk: A Hippocratic Oath for Data Science | slides | pdf
• June 21 | PyData London | Python for Image Understanding: Deep Learning with Convolutional Neural Nets | slides | pdf | video
• June 10 | Stockholm NLP Meetup | Explore Data: Data Science + Visualization | slides
• May 12 at PyCon Sweden | Conference | Deep Learning as a Cat/Dog Detector | slides
| pdf | video | ipython notebook (code)
• April 28 at KTH | Guest Lecture DD2476 | Deep Learning for Information Retrieval | slides
• April 20 at Sthlm Deep Learning Meetup | Deep Learning: a (non-techy) birds eye view | abstract | slides | video
• April 20 at Sthlm Deep Learning Meetup | Graphs and Deep Learning | abstract | slides | video
• April 11 at KTH | Vision/Language Deep Learning reading group | presenting Hill et al. Learning to understand phrases by embedding the dictionary, 2015 | slides
• February 20 at KTH | Vision/Language Deep Learning reading group | presenting Socher et al. Zero-shot learning through cross-modal transfer, ICLR 2013 | slides
• January 23 at KTH, CVAP | Seminar | Visual-Semantic Embeddings: some thoughts on Language | abstract | slides
• January 22 at Sthlm NLP Meetup | NLP and Deep Learning: introduction - Session I | abstract | slides | video | IPython Notebooks (code)

2014

• December 4 at KTH | Guest Lecture in course: Language Technology | Deep Learning for NLP: An Introduction to Neural Word Embeddings | slides
• October 21 at Sics | Deep Learning & NLP: Graphs to the Rescue! | abstract | slides
• May 14 at KTH | Recommender Systems, Matrices and Graphs | abstract | slides

Investigation of Different Nets and Layers

VIS_bvlc_googlenet_allLayers
VIS_googlenet_places205_allLayers
VIS_finetune_flickr_style_allLayers
VIS_alexnet_places_allLayers
VIS_oxford102flowers_allLayers

Painting Experiments

(mainstream news: gizmodo, twitter)

Video Experiments

(mainstream news: wired, businessinsider, gizmodo)

Audio Experiments

Contact information:

Email:	roelof [at] csc.kth.se
PGP-key:	pub (0xB4FDE472)
Phone:	+46 76 317 38 48
Post Adress:	KTH CSC, 100 44 Stockholm
Visitors Adress:	Lindstedtsvägen 5, plan 5, room 4540