From awd at cs.cmu.edu Fri Mar 19 18:06:49 2010 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Fri, 19 Mar 2010 18:06:49 -0400 Subject: [Research] Lab meeting: Wednesday March 24 Message-ID: <4BA3F579.9010809@cs.cmu.edu> Time: The First Wednesday of Spring, 12:00noon Place: NSH 1507 (Karen W. to confirm that) Speaker: David Anqi Cui Title: Active Learning Algorithms for Fast Drug Discovery Abstract: In this talk, I will show my work during the past six months. The project is based on drug discovery which usually costs a lot of time and money. Chemists have to do many experiments to screen useful compounds. In this paper, we present active learning algorithms for fast drug discovery. The algorithms decrease the number of experiments required to find out the best performance compounds among plenty of possible trials. The problem is a traditional exploration vs. exploitation dilemma and our approach is based on the multi-armed bandit problem and other function approximators. We propose the expected improvement estimation as a method to measure the unknown compounds. Some traditional models including UCB algorithms, Gaussian process, regression trees and so on are also used for this problem. Our results show that the algorithms present in this paper significantly raise the best performance of compounds found within a certain number of picks. The number of picks needed to first discover the best compound are also reduced to about half of random method's cost. See you all there! From awd at cs.cmu.edu Sun Mar 21 09:35:55 2010 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Sun, 21 Mar 2010 09:35:55 -0400 Subject: [Research] Lab meeting: Wednesday March 24 In-Reply-To: <4BA3F579.9010809@cs.cmu.edu> References: <4BA3F579.9010809@cs.cmu.edu> Message-ID: <4BA620BB.7000802@cs.cmu.edu> The meeting has been moved to HBH 6501. Other parts of the plan remain unchanged. Artur Artur Dubrawski wrote: > Time: The First Wednesday of Spring, 12:00noon > > Place: NSH 1507 (Karen W. to confirm that) > > Speaker: David Anqi Cui > > Title: Active Learning Algorithms for Fast Drug Discovery > > Abstract: > > In this talk, I will show my work during the past six months. The > project is based on drug discovery which usually costs a lot of time > and money. Chemists have to do many experiments to screen useful > compounds. In this paper, we present active learning algorithms for > fast drug discovery. The algorithms decrease the number of experiments > required to find out the best performance compounds among plenty of > possible trials. The problem is a traditional exploration vs. > exploitation dilemma and our approach is based on the multi-armed > bandit problem and other function approximators. We propose the > expected improvement estimation as a method to measure the unknown > compounds. Some traditional models including UCB algorithms, Gaussian > process, regression trees and so on are also used for this problem. > Our results show that the algorithms present in this paper > significantly raise the best performance of compounds found within a > certain number of picks. The number of picks needed to first discover > the best compound are also reduced to about half of random method's > cost. > > See you all there! > > _______________________________________________ > Research mailing list > Research at autonlab.org > https://www.autonlab.org/mailman/listinfo/research > > From awd at cs.cmu.edu Sun Mar 21 09:36:52 2010 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Sun, 21 Mar 2010 09:36:52 -0400 Subject: [Research] Lab meeting: Wednesday March 24 In-Reply-To: <4BA620BB.7000802@cs.cmu.edu> References: <4BA3F579.9010809@cs.cmu.edu> <4BA620BB.7000802@cs.cmu.edu> Message-ID: <4BA620F4.8010604@cs.cmu.edu> Of course I meant GHC not HBH. I must be getting older or something :) Artur Dubrawski wrote: > The meeting has been moved to HBH 6501. > Other parts of the plan remain unchanged. > > Artur > > Artur Dubrawski wrote: >> Time: The First Wednesday of Spring, 12:00noon >> >> Place: NSH 1507 (Karen W. to confirm that) >> >> Speaker: David Anqi Cui >> >> Title: Active Learning Algorithms for Fast Drug Discovery >> >> Abstract: >> >> In this talk, I will show my work during the past six months. The >> project is based on drug discovery which usually costs a lot of time >> and money. Chemists have to do many experiments to screen useful >> compounds. In this paper, we present active learning algorithms for >> fast drug discovery. The algorithms decrease the number of experiments >> required to find out the best performance compounds among plenty of >> possible trials. The problem is a traditional exploration vs. >> exploitation dilemma and our approach is based on the multi-armed >> bandit problem and other function approximators. We propose the >> expected improvement estimation as a method to measure the unknown >> compounds. Some traditional models including UCB algorithms, Gaussian >> process, regression trees and so on are also used for this problem. >> Our results show that the algorithms present in this paper >> significantly raise the best performance of compounds found within a >> certain number of picks. The number of picks needed to first discover >> the best compound are also reduced to about half of random method's >> cost. >> >> See you all there! >> >> _______________________________________________ >> Research mailing list >> Research at autonlab.org >> https://www.autonlab.org/mailman/listinfo/research >> >> > > From awd at cs.cmu.edu Mon Mar 22 20:08:32 2010 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Mon, 22 Mar 2010 20:08:32 -0400 Subject: [Research] T-Cube Web Interface on Canadian TV [Fwd: Fwd: CBC story on RTBP] Message-ID: <4BA80680.1090208@cs.cmu.edu> -------------- next part -------------- An embedded message was scrubbed... From: Nuwan Waidyanatha Subject: Fwd: CBC story on RTBP Date: Tue, 23 Mar 2010 07:59:34 +0800 Size: 5696 URL: From krw at andrew.cmu.edu Tue Mar 23 15:08:47 2010 From: krw at andrew.cmu.edu (Karen Widmaier) Date: Tue, 23 Mar 2010 15:08:47 -0400 Subject: [Research] Lab meeting: Wednesday March 24 In-Reply-To: <4BA620F4.8010604@cs.cmu.edu> References: <4BA3F579.9010809@cs.cmu.edu> <4BA620BB.7000802@cs.cmu.edu> <4BA620F4.8010604@cs.cmu.edu> Message-ID: <02ff01cacabc$43bef190$cb3cd4b0$@cmu.edu> Just a reminder regarding the room for tomorrow's meeting will be GHC 6501 at noon. -----Original Message----- From: research-bounces at autonlab.org [mailto:research-bounces at autonlab.org] On Behalf Of Artur Dubrawski Sent: Sunday, March 21, 2010 9:37 AM To: research at autonlab.org Subject: Re: [Research] Lab meeting: Wednesday March 24 Of course I meant GHC not HBH. I must be getting older or something :) Artur Dubrawski wrote: > The meeting has been moved to HBH 6501. > Other parts of the plan remain unchanged. > > Artur > > Artur Dubrawski wrote: >> Time: The First Wednesday of Spring, 12:00noon >> >> Place: NSH 1507 (Karen W. to confirm that) >> >> Speaker: David Anqi Cui >> >> Title: Active Learning Algorithms for Fast Drug Discovery >> >> Abstract: >> >> In this talk, I will show my work during the past six months. The >> project is based on drug discovery which usually costs a lot of time >> and money. Chemists have to do many experiments to screen useful >> compounds. In this paper, we present active learning algorithms for >> fast drug discovery. The algorithms decrease the number of experiments >> required to find out the best performance compounds among plenty of >> possible trials. The problem is a traditional exploration vs. >> exploitation dilemma and our approach is based on the multi-armed >> bandit problem and other function approximators. We propose the >> expected improvement estimation as a method to measure the unknown >> compounds. Some traditional models including UCB algorithms, Gaussian >> process, regression trees and so on are also used for this problem. >> Our results show that the algorithms present in this paper >> significantly raise the best performance of compounds found within a >> certain number of picks. The number of picks needed to first discover >> the best compound are also reduced to about half of random method's >> cost. >> >> See you all there! >> >> _______________________________________________ >> Research mailing list >> Research at autonlab.org >> https://www.autonlab.org/mailman/listinfo/research >> >> > > _______________________________________________ Research mailing list Research at autonlab.org https://www.autonlab.org/mailman/listinfo/research