Upcoming Workshops & Hackathons:

UPDATE: Due to COVID19, our second set of workshops/hackathons will likely be delayed; thus, we are putting together an online module about DeepLabCut: Summer 2020 DLC Course

March 5th & 6th, 2020: Hackathon at Harvard University was a great success. We worked primarily on hardware integration for real-time deeplabcut. Special guests Gonçalo Lopes, Jonny Saunders and our own Gary Kane worked on affordable open-source solutions for behavioral feedback using deeplabcut (read more about them below).

There will be three software releases:

Hackathon 2020 Special Guests!

Evan Shelhamer

Evan Shelhamer

Jonny Saunders

Jonny Saunders

Gonçalo Lopes

Gonçalo Lopes

Jordan Salvi

Jordan Salvi

Gonçalo Lopes is a software engineer turned neuroscientist, with a background of applied research in augmented reality and robotics. Gonçalo completed his PhD from the Champalimaud Neuroscience Programme in 2016, working with Adam Kampff and Joe Paton to understand the role of motor cortex in the robust control of movement. Along the way, he developed the Bonsai visual programming language (https://bonsai-rx.org) as an accessible platform for scientific experimentation, which has been increasingly adopted in neuroscience to rapidly prototype rich and dynamic closed-loop environments. The goal of Bonsai is also to facilitate technology integration, allowing easy combination of open-source hardware and software used in neuroscience such as open-ephys, miniscopes, behavior cameras, photometry systems, or arduinos and other microcontrollers for controlling behavior environments. Gonçalo is now the founder of NeuroGEARS Ltd (https://neurogears.org/), where he continues to develop and use Bonsai to empower scientists, students, artists and other enthusiasts to take control of real-time digital information technology to realize their scientific fantasies. Gonçalo teaches hands-on workshops on Bonsai, and has running collaborations with a number of research groups and organizations around the world to promote the development of widely accessible neuroscience research tools. At this workshop, Gonçalo will be working to develop a Bonsai package for DeepLabCut, so its tracking technology can be easily combined with other hardware and software elements to design new kinds of behavior experiments.

Evan Shelhamer is an experimenter, engineer, and evangelist for DIY deep learning. He completed his PhD at UC Berkeley, advised by Trevor Darrell, and has now taken his science to Massachusetts as an Adobe research scientist and MIT collaborator. His research is focused on reconciling visual structures like locality and scale with end-to-end learning. Evan believes in open science and open tooling for research and engineering, and that is why he was the lead developer of the Caffe deep learning framework from version 0.1 to 1.0. He still engages in open sourcery when he can, like for this workshop! At the workshop, Evan will hack on optimization and model uncertainty by incorporating a fuller sort of likelihood into the training of the DeepLabCut fully convolutional networks. Ask him about brewing coffee, brewing community, and taking classic ideas and making them differentiable. He likes his coffee black, models deep, and descent steep.

Jonny Saunders is a systems neuroscience PhD student at the University of Oregon. They study computational mechanisms of complex sounds in auditory cortex with Michael Wehr. Currently they are working on grounding theoretical models of speech processing in neurophysiological data, investigating the unexpected role of extracellular protein matrices in auditory cortical plasticity, and a project about the neural computation of hip-hop that you can only faintly make out as billows of smoke and vague mumbling noises coming from behind a veil of mystery. They recently released Autopilot (https://auto-pi-lot.com), a Python framework that leverages distributed swarms of Raspberry Pis to perform the next generation of complex behavioral neuroscience experiments. They are looking for collaborators to test and develop Autopilot, as well as labs that are willing to host them for a few weeks in this coming summer to develop an experiment for them in order to build out Autopilot's capabilities. At this workshop, Jonny will be working on integrating DeepLabCut into a realtime image-processing pipeline in Autopilot so that tracked points can be used as parameters and triggers in behavioral experiments.
Likes: messy, intractable scientific problems; unlikely cross-disciplinary collaboration; that ancient geocities page that you can't delete because you forgot the password in 2002. Dislikes: bad science; opaque, entrenched power; people googling their old screennames.

Jordan Salvi is a software engineer and masters student working with Backyard Brains engage k12 teachers and students in citizen science projects with neuroscience researchers. Our goal is to develop a web based interface for DeepLabCut, so that students can assist with the annotation of video frames for neuroscientists.  This interface will provide a useful tool to automate new preps for scientists, as well as allowing students to experience how Deep Neural Networks work within behavioural science. 

WORKSHOPS

Jan 16th & 17th, 2020: new users (Harvard University)

  • The application is now closed! (closes Dec 1st)

  • There will be another workshop in Europe (CH) in Fall 2020.

Past Events

DLC_workshop2019-01.png

In 2019 we hosted a workshop for new users of DeepLabCut at Harvard and other locations:

  • CSHL July 2019

  • MBL June 2019: Marine Biological Laboratory, Woods Hole, MA

  • June 29-July 6, 2019: CVSS 2019: Black Forest, Germany (http://orga.cvss.cc/)

  • Jan 28-29 2019: Rowland Institute, Cambridge, MA*

  • Jan 7th, 2019: Marine Biological Laboratory, Woods Hole, MA (locally organized by Dr. Yisrael Schnytzer)

  • Dec 13-14th, 2018: Nencki Institute of Experimental Biology, Warsaw (locally organized by Mateusz Kostecki)

  • Dec 10-11th, 21018: Ludwig-Maximilians-University, Munich (locally organized by Kate Stynik)

More info