Learning from Label Proportions (LLP) Online ind Onl ine - PowerPoint PPT Presentation

Apr 09, 2023 •196 likes •239 views

(Almost) No Label No Cry Giorgio Patrini with R.Nock, P.Rivera, T.Caetano Learning from Label Proportions (LLP) Online ind Onl ine individual r ividual recor ecords ds Per Percent cent unemployment unemployment Input : unlabeled data

(Almost) No Label No Cry Giorgio Patrini with R.Nock, P.Rivera, T.Caetano Learning from Label Proportions (LLP) Online ind Onl ine individual r ividual recor ecords ds Per Percent cent unemployment unemployment Input : unlabeled data Input : label proportions Output : predictor of individual unemployment How likely Alice is unemployed given only her online behavior
(Almost) No Label No Cry Giorgio Patrini with R.Nock, P.Rivera, T.Caetano Our Solution Def, Altun&Smola ’06: the mean operator µ = 1 /m P m i =1 y i x i Thm: is su ffi cient for the label variable for most Proper Losses: µ proper-loss = loss w/o labels ( θ ) − 1 2 < θ , µ > • Quadrianto et al. ‘09, • Our relaxation: homogeneity assumption : “The more similar the counties, the more “Unemployed people in all the counties similar the online behavior of the behave online in the same way” unemployed people”
(Almost) No Label No Cry Giorgio Patrini with R.Nock, P.Rivera, T.Caetano Results • Finite sample approximation bounds for the resulting classifier (do not hold for previous approaches) • First generalization result for LLP , based on Rademacher complexity #proportions/#instances “more supervised”

Download Presentation

Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

Comparing two proportions Beginning Bayes in R Learning about many parameters Chapters 2-3

BEGINNING BAYES IN R Comparing two proportions Beginning Bayes in R Learning about many parameters Chapters 2-3 : single parameter (one proportion or one mean) Chapter 4 : multiple parameters Two proportions from independent

981 views • 69 slides

Learning Context-dependent Label Permutations for Multi-label Classification Jinseok Nam Amazon

Learning Context-dependent Label Permutations for Multi-label Classification Jinseok Nam Amazon Alexa AI Joint work with Young-Bum Kim, Eneldo Loza Menca, Sunghyun Park, Ruhi Sarikaya and Johannes Frnkranz Mu Multi-lab label el Clas

742 views • 13 slides

AG! Blue Label Bench-top Reactor 1 Product line up Platinum Label Gold Label Blue Label Blue

AG! Blue Label Bench-top Reactor 1 Product line up Platinum Label Gold Label Blue Label Blue Label Plug & Play : Ease of Use Gold Label Flexible and Easy Upgrades and Customization Platinum Label Custom Chemical Engineering Solutions 2

591 views • 22 slides

SIGML Department of CSE IIT Kanpur Special Interest Group in Machine Learning Classification

Multi-label Learning Trees, Embeddings, and much more! Purushottam Kar SIGML Department of CSE IIT Kanpur Special Interest Group in Machine Learning Classification Paradigms Pick one Pick one Pick all applicable Label 1 Label 1 Label 1

839 views • 32 slides

Blue Label Pilot-plant Reactor 1 Product Line-up Platinum Label Gold Label Blue Label Blue

Blue Label Pilot-plant Reactor 1 Product Line-up Platinum Label Gold Label Blue Label Blue Label Plug & Play : Ease of Use Gold Label Flexible and Easy Upgrades and Customization Platinum Label Custom Chemical Engineering Solutions 2 Two

377 views • 24 slides

Machine Learning for Computational Linguistics May 24, 2016 . ltekin, label of an unknown

Machine Learning for Computational Linguistics May 24, 2016 . ltekin, label of an unknown label. In the example: Supervised learning: classifjcation PCA Hierarchical clustering Mixture densities K-means Clustering Introduction

103 views • 7 slides

Small area estimation of proportions of Small area estimation of proportions of Arsenic affected

Small area estimation of proportions of Small area estimation of proportions of Arsenic affected wells in Bangladesh Arsenic affected wells in Bangladesh By Sanghamitra Pal West Bengal State University, India (Joint work with Prof. Partha

495 views • 32 slides

Presentation of the label Certicold WHY A CERTICOLD LABEL? A European conformity label For

Presentation of the label Certicold WHY A CERTICOLD LABEL? A European conformity label For European users and manufacturers of professional equipment for the cold chain. Created by professionals for professionals Managed by an INDEPENDENT

219 views • 19 slides

Jointly Learning to Label Sentences and Tokens Marek Rei Anders

Jointly Learning to Label Sentences and Tokens Marek Rei Anders Sgaard 1/12 Task 1: Sentence Classification Error Detection It was so long time to wait in the theatre . I like to playing the guitar and sing

297 views • 12 slides

Reducing Label Cost by Combining Feature Labels and Crowdsourcing Combining Learning Strategies

Reducing Label Cost by Combining Feature Labels and Crowdsourcing Combining Learning Strategies to Reduce Label Cost 7/2/2011 Jay Pujara jay@cs.umd.edu Ben London blondon@cs.umd.edu Lise Getoor getoor@cs.umd.edu University of Maryland,

524 views • 16 slides

Multi-label Learning Approaches for Music Instrument Recognition Eleftherios Spyromitros - Xioufis

Analyzing the data Engineering the input Exploring multi-label approaches Engineering the output Conclusions Multi-label Learning Approaches for Music Instrument Recognition Eleftherios Spyromitros - Xioufis , Grigorios Tsoumakas and Ioannis

658 views • 18 slides

Unit 5: Inference for categorical variables Lecture 1: Inference for proportions Statistics 101

Unit 5: Inference for categorical variables Lecture 1: Inference for proportions Statistics 101 Thomas Leininger June 12, 2013 Many research questions involve proportions Who will win the election?

350 views • 23 slides

Poisson Learning: Graph-based semi-supervised learning at very low label rates Jeff Calder 1 ,

Poisson Learning: Graph-based semi-supervised learning at very low label rates Jeff Calder 1 , Brendan Cook 1 , Matthew Thorpe 2 , and Dejan Slep cev 3 1 School of Mathematics, University of Minnesota 2 Department of Mathematics, University of

1.23k views • 46 slides

IETF 78 TPA-Label for ADSP DKIM Third-Party Authorization Label draft-otis-dkim-tpa-label By

IETF 78 TPA-Label for ADSP DKIM Third-Party Authorization Label draft-otis-dkim-tpa-label By Douglas Otis & Daniel Black Acceptance w/o Author Domain Signature and ADSP Reject or discard discardable w/o A-D sig breaks mailing lists

473 views • 10 slides

Which data do we need for training? Domain Adaption and Learning under Label Noise Franz

56 th Photogrammetric Week Stuttgart 20170913 Which data do we need for training? Domain Adaption and Learning under Label Noise Franz Rottensteiner Institute of Photogrammetry and GeoInformation Leibniz Universitt Hannover

1.21k views • 53 slides

Work on Multi-label Classification Jesse Read Supervised by Bernhard Pfahringer

Work on Multi-label Classification Jesse Read Supervised by Bernhard Pfahringer jmr30@cs.waikato.ac.nz Machine Learning Group University of Waikato Hamilton New Zealand Work on Multi-label Classification p. 1/1 Outline Multi-label

400 views • 29 slides

Welcome to the Cyber Risk Insights Conference! Welcoming Remarks Rebecca Bole EVP &

Welcome to the Cyber Risk Insights Conference! Welcoming Remarks Rebecca Bole EVP & Editor-in-Chief Advisen Opening Remarks Jeremy Barnett Senior Vice President NAS Insurance Services Keynote Address Suzanne Spaulding Senior

1.14k views • 64 slides

Logics for Reasoning about Logics for Reasoning about Quantum Information: Quantum Information:

Logics for Reasoning about Logics for Reasoning about Quantum Information: Quantum Information: A Dynamic-Epistemic Perspective A Dynamic-Epistemic Perspective Sonja Smets Sonja Smets University of Amsterdam University of Amsterdam Based

666 views • 43 slides

Outline 1. Optical Quantum Computing 101 2. Where we are 3. Some theoretical magic 4.

Progress Toward Scalable Linear Optical Quantum Computing Paul Kwiat Outline 1. Optical Quantum Computing 101 2. Where we are 3. Some theoretical magic 4. Technology Development -- where we need to get to Sources Detectors

685 views • 53 slides

Lecture 13 CSE 260 Parallel Computation (Fall 2015) Scott B. Baden Message Passing Stencil

Lecture 13 CSE 260 Parallel Computation (Fall 2015) Scott B. Baden Message Passing Stencil methods with message passing Announcements Weds office hours changed for the remainder of quarter: 3:30 to 5:30 Scott B. Baden / CSE 260, UCSD

629 views • 35 slides

Week 8, Lectures 1 & 2 : Fixed-, Random-, and Mixed-Effects models 1. The repeated measures

Week 8, Lectures 1 & 2 : Fixed-, Random-, and Mixed-Effects models 1. The repeated measures design, where each of n Ss is measured k times, is a popular one in Psych. We approach this design in 2 ways: 1. As a generalisation of the paired

728 views • 61 slides

Workshop 9.3a: Randomized block designs Murray Logan November 23, 2016 Table of contents 1

. . . . . . . . . . . . . . . . . . . . . . . . . . -1- Workshop 9.3a: Randomized block designs Murray Logan November 23, 2016 Table of contents 1 Randomized Block (RCB) designs 1 2 Worked Examples 12 1.

383 views • 12 slides

The Linear Mixed-Effects Model The R Statistical Computing Environment Basics and Beyond The

The Linear Mixed-Effects Model The R Statistical Computing Environment Basics and Beyond The Laird-Ware form of the linear mixed model: Mixed-Effects Models y ij = 1 + 2 x 2 ij + + p x pij + b 1 i z 1 ij + + b qi z qij

399 views • 4 slides

CS 5150 So(ware Engineering 12. System Architecture William Y. Arms Design The requirements

Cornell University Compu1ng and Informa1on Science CS 5150 So(ware Engineering 12. System Architecture William Y. Arms Design The requirements describe the funcCon of a system as seen by the client. For a given set of requirements, the

782 views • 28 slides