Daniel

Research scientist at NEC Labs Europe. Computer vision and machine learning expert.

[Mathias Niepert @Mniepert]

Our group at NEC Labs Europe has several talks and posters at @akbc_conf this year. Come speak to @caro__lawrence @Daniel_Onoro and Bhushan at our booth or the posters. They are super nice!

[Mathias Niepert @Mniepert]

#emnlp2018 paper on temporal KB completion using LSTMs to encode time expressions such as "president since 2008" as vectors. These are combined with KB embedding methods https://t.co/hgBOBTUkGe We also provide 4 datasets with temporal facts https://t.co/8BXpK9ySji @sebdumancic

[Daniel @Daniel_Onoro]

I will be in Newcastle upon Tyne presenting our work "Learning Short-Cut Connections for Object Counting" at the BMVC 2018. Visits to my poster are wellcome :) #BMVC2018

[Daniel @Daniel_Onoro]

I am in KDD 2018 Deep Learning Day presenting our work "Representation Learning for Visual-Relational Knowledge Graphs". #KDD2018

[Mathias Niepert @Mniepert]

KBLRN combines logical, learned, and numerical features (geolocation, age, etc.) into joint model. Probabilistic product of experts trained with negative sampling. Works better than more complex methods for combining these feature types. Come visit Alberto and our poster @uai2018

[Mathias Niepert @Mniepert]

We won best poster award at @midl_amsterdam! Contextual hourglass networks create shortcut connections between layers with different spatial dimensions, tying movement of two filter banks. Applicable to U-Net, V-Net, etc. Simple but elegant. Congrats @Daniel_Onoro! #midl2018

[ipfconline @ipfconline1]

Artificial Intelligence in #Fintech
👉A Compilation of Great Research Papers

https://t.co/ylbeQx6jBQ [by @slavakurilyak v/ @produvia]
#AI #MachineLearnig #DeepLearning #NeuralNetworks #NLP #Algorithms #CustServ #Cybersecurity #Marketing

Cc @DeepLearn007 @evankirstel @guzmand

[evolvingstuff @evolvingstuff]

BrainSlug: Transparent Acceleration of Deep Learning Through Depth-First Parallelism
https://t.co/fM5PF8LzrN

[Mathias Niepert @Mniepert]

Knowledge graph datasets based on DBpedia and Yago with images, numerical features, and sameAs links to FB15k. Benchmarks for multi-modal KB completion methods. https://t.co/8BXpK9ySji