Department of Computer Science | Institute of Theoretical Computer Science | CADMO

Theory of Combinatorial Algorithms

Prof. Emo Welzl and Prof. Bernd Gärtner

Mittagsseminar (in cooperation with A. Steger, D. Steurer and B. Sudakov)

Mittagsseminar Talk Information

Date and Time: Tuesday, October 30, 2012, 12:15 pm

Duration: 30 minutes

Location: OAT S15/S16/S17

Speaker: Hafsteinn Einarsson

One shot learning

How can a toddler learn a word by hearing it once? The process of how this works in the brain is ill understood and it touches upon the subject of the fine balance between cortical plasticity and stability.

It is widely accepted that information in the brain is encoded with sets of neurons spiking almost simultaneously. When learning associations between two sets of neurons by increasing the connection strength between them we run the risk of overwriting old information. To overcome this, traditional neuron simulations have assumed slow change in connection weights to preserve information in the network. With this method we need several rounds of learning to consolidate information which contradicts the fact that we are able to learn in one shot.

In this talk I present and give a short analysis of a stochastic learning rule which gives the best of both worlds. We can learn in one shot while not ruining the structure of the network.

Upcoming talks     |     All previous talks     |     Talks by speaker     |     Upcoming talks in iCal format (beta version!)

Previous talks by year:   2024  2023  2022  2021  2020  2019  2018  2017  2016  2015  2014  2013  2012  2011  2010  2009  2008  2007  2006  2005  2004  2003  2002  2001  2000  1999  1998  1997  1996  

Information for students and suggested topics for student talks

Automatic MiSe System Software Version 1.4803M   |   admin login