This is not the document you are looking for? Use the search form below to find more!

Report home > Others

Fixing the program my computer learned: End-user debugging of machine-learned programs

0.00 (0 votes)
Document Description
Fixing the program my computer learned: End-user debugging of machine-learned programs
File Details
  • Added: April, 25th 2011
  • Reads: 146
  • Downloads: 0
  • File size: 2.00mb
  • Pages: 30
  • Tags:
  • content preview
Submitter
  • Name: desi
Embed Code:

Add New Comment




Related Documents

End User Computing – An User Centric Process

by: hclisd, 2 pages

In the field of computing “end user computing” would refer to a structure in which the non-programmers can set up working applications. It comprise group of approaches in computing that ...

End User Computing Solutions to Collaborate Efficiently

by: hclisd, 1 pages

In today’s highly competitive world, it is essential for end-user to have the right tools to work and collaborate efficiently. That is not all; IT needs to secure and manage the endpoints that ...

Stimulating IT Tools to Accessorize the End User’s Flair

by: hclisd, 1 pages

IT has been a generous facilitator and provided us with infinite advantages. It has made life much easier with the conversion of complex tasks into simple processes. It has empowered businesses with ...

Technology Controls in Business - End User Computing

by: harry, 10 pages

End User Computing Technology Controls in Business Renetta Ho-Antonio PMCP, ERM, CISM Introduction The Use of Spreadsheets: As it relates to Section 404 of the ...

Tobacco Price Support: An Overview of the Program

by: shinta, 7 pages

About 94% of U.S. tobacco production is flue-cured and burley (both being cigarette tobacco types). These crops are particularly important to the agriculture of North Carolina (where ...

Summer of Code at the New Zealand Computer Society Evening With industry 2008

by: cerys, 29 pages

Summer of Code at the New Zealand Computer Society Evening With industry 2008

Global social policy addressing the internal brain drain of medical doctors in thailand - the story and lesson learned

by: beert, 5 pages

Global social policy addressing the internal brain drain of medical doctors in thailand - the story and lesson learned

How To Compute The Best Fit Straight Line To A Set Of Data?

by: madison, 5 pages

How-To Compute the Best-Fit Straight Line to a Set of Data Page 1 How-To Compute the Best Fit Straight Line to a Set of Data ...

User Evaluation of Dublin Core Metadata in Image Collections

by: pietronella, 52 pages

User Understanding of Dublin Core Metadata in Digital Image CollectionsKathleen FearJuly 9, 2009IntroductionMethodologyFindingsDiscussion and ...

Imports in the Washington State Economy: Importance and Regional Effects of Import Liberalization

by: shinta, 44 pages

This paper focuses on the import side of a regional economy quantifying the economic impact of import levels and trade liberalization. An innovation represents the linkage of a regional ...

Content Preview
Fixing the program my computer learned: End-user debugging of machine-learned programs Dr Simone Stumpf City University London Simone.Stumpf.1@city.ac.uk Bio 1996 BSc, Comp Sci w/ Cog Sci, UCL 2001 PhD Comp Sci, UCL 2001 - 2004 Research Fellow, UCL 2004 - 2007 Research Manager, Oregon State (OSU) 2007 - 2009 UX Architect, White Horse 2008 - present Asst Professor (Senior Research), OSU 2009 - present Lecturer, City University London 2 What are machine-learned programs? •! Systems that “predict” –! Spam filters, “smart desktops”, web page recommendations •! Learn from and adapt to user after deployment •! Probabilistic machine learning algorithms •! Resulting behaviour is a program How do you debug a program that was written by a machine instead of a person? Especially when you don’t know much about programming and are working with a program you can’t even see? 3 A quick machine learning detour… “Simple” algorithm like Naïve Bayes –! Have input (features) and outputs (labels or classes) –! From training data they learn a function: weight*input = class –! As they further learn weights are changed eg. spam filters (bag-of-words approach) –! take all words appearing in the training data as features –! throw out stop words (a, the, ?) –! do stemming (walking, walked = walk) –! learn how prevalent certain words are in spam messages –! use that function to predict whether new email message is spam 4 Current debugging approach Based on your interest in: We recommend: ! ! ! " 5 Problems and opportunities for end users •! Are not machine learning experts or programmers •! Only they can fix if incorrect behaviour occurs –! Cannot inspect source code –! Can only observe results at run-time –! Can usually only give more training examples to influence future behaviour –! Need to provide lots of training data to change behaviour •! Much richer knowledge could be exploited •! Could increase usability and trust How can the program communicate its reasoning to the end user? How could the user talk back? 6 Formative study •! Enron email dataset folders (farmer-d): Personal, Resume, Bankrupt, Enron News (122 messages) •! Lo-fi prototypes with explanations –! Rule-based –! Similarity-based –! Keyword-based •! 13 participants, talk-aloud 7 Explanations by ML program Simplified yet faithful Concrete •! Rule-based best understood but no clear overall preference •! Serious understandability problems with Similarity-based •! Negative keyword list with keyword-based problematic (negative weights) Matters if they they think reasoning is sound and it is communicated clearly, word choices important 8 What does the user tell the program? •! Select different features (53%) –! It should put email in ‘Enron News’ if it has the keywords “changes” and “policy”. •! Adjust weights (12%) –! The second set of words should be given more importance. •! Parse/extract in different way (10%) –! I think that it should look for typos in the punctuation for indicators toward ‘Personal’. •! Employ feature combinations (5%) –! I think it would be better if it recognized a last and a first name together. •! Use relational features (4%) –! This message should be in ‘EnronNews’ since it is from the chairman of the company. 9 What knowledge do they use? •! Commonsense (36%) –! “Qualifications” would seem like a really good Resume word, I wonder why that’s not down here. •! English (30%) –! Does the computer know the difference between “resumé” and “resume”? •! Domain (15%) –! Different words could have been found in common like … “Ken Lay”. 10

Download
Fixing the program my computer learned: End-user debugging of machine-learned programs

 

 

Your download will begin in a moment.
If it doesn't, click here to try again.

Share Fixing the program my computer learned: End-user debugging of machine-learned programs to:

Insert your wordpress URL:

example:

http://myblog.wordpress.com/
or
http://myblog.com/

Share Fixing the program my computer learned: End-user debugging of machine-learned programs as:

From:

To:

Share Fixing the program my computer learned: End-user debugging of machine-learned programs.

Enter two words as shown below. If you cannot read the words, click the refresh icon.

loading

Share Fixing the program my computer learned: End-user debugging of machine-learned programs as:

Copy html code above and paste to your web page.

loading