Anthony Platanios

8221 Gates Hillman Center
5000 Forbes Ave
Pittsburgh, PA 15213
I am a PhD student in the Machine Learning Department of the School of Computer Science at Carnegie Mellon University. My advisor is Tom Mitchell and I work on Never-Ending Learning. My current research is motivated by the fact that real-world problems require integrating multiple, distinct modalities of information (e.g., image, audio, language, etc.) in ways that machine learning models cannot currently handle well. Most deep learning approaches are not able to utilize information learned from solving one problem to directly help in solving another. They are also not capable of never-ending learning, failing on problems that are dynamic, ever-changing, and not fixed a priori, which is true of problems in the real world due to the dynamicity of nature. With my research, I aim to bridge the gap between UTCs, deep learning, and never-ending learning, by proposing neural cognitive architectures (NCAs) that are inspired by human cognition and that can learn to continuously solve multiple problems that can grow in number over time, across multiple distinct perception and action modalities, and from multiple noisy sources of supervision combined with self-supervision. Their experience from learning to solve past problems can also be leveraged to learn to solve future ones. If you are interested to read more about NCAs, my thesis proposal would be a good place to start. Throughout my PhD I have also worked on multiple other projects related to artificial intelligence and machine learning.

Before I joined CMU, I graduated with an M.Eng. in Electrical and Electronic Engineering from Imperial College London. For my Master's thesis I proposed a way to use topic modelling methods in order to perform human motion classification.

news

May 3, 2019

Organizing the Adaptive & Multi-Task Learning workshop at ICML 2019.

May 2, 2019

Organizing the Learning with Limited Labeled Data workshop at ICLR 2019.

Oct 25, 2018

Released a new version of TensorFlow Scala that finally introduces type-safety throughout the graph construction process including autodiff.

Dec 1, 2017

Attending NIPS 2017 and presenting some of our work on estimating accuracy.

Nov 20, 2017

Organizing the Learning with Limited Labeled Data workshop at NIPS 2017.

May 26, 2017

Open sourced the TensorFlow Scala library.

Jul 29, 2016

Received the Carnegie Mellon University Presidential Fellowship.