Deep learning (DL) has in recent years become the state of the art technique for a wide variety of computer vision and NLP problems, in addition to breakthrough results in everything from drug discovery, to atomic physics, to dermatology. The scientific committee of the inaugural Data Institute Conference is pleased to announce this call, to be chaired by Jeremy Howard, former CEO of Enlitic and past president of Kaggle, current Distinguished Scholar in Deep Learning at USF and co-founder of fast.ai.
Topics can cover any development related to deep learning, broadly defined, such as:
As an academic conference, the committee is looking for technical talks, but are also trying to make the conference more accessible than most, so that more people can enjoy a wider range of talks. Therefore, we are asking people to spend some time thinking about how best to present their topic to a technical audience of people who may not necessarily be experts in the specific area of the talk. The goal here is to increase the level of collaboration between academia and industry.
Abstracts should not be longer than 1 page and include a statement of the problem, methods and approach as well as a summary of results, and no more than 5 relevant citations. Abstracts will be reviewed by the deep learning scientific committee chaired by Jeremy Howard. Experimental results on "real world" datasets are particularly encouraged. We are looking for excellent presenters, not just excellent researchers, so where possible please also submit a video of up to 5 minutes briefly explaining your work and its context (videos need not be high quality productions---just a short talk into your webcam is sufficient), or provide a link to one of your previous talks that has been recorded.
As part of the submission procedure authors are asked to mark conflicts of interest with Program Committee members. A poster author or contributed speaker has a conflict of interest with a Program Committee member if any of the following hold: