SlideShare a Scribd company logo
1 of 18
CVPR 2019 oral
Laso: Label-Set Operations Networks for Multi-label Few-shot Learning
Amit Alfassy, Leonid Karlinsky, Amit Aides
IBM Research AI
Haifa, Israel
Source – https://slideplayer.com/slide/17422890/
思想自由 兼容并包
Note – Paper is presented from the point of view of Few shot learning for multi-
label samples
思想自由 兼容并包
Methodology
Combine pairs of given examples in feature space, so that the resulting synthesized feature vectors will
correspond to examples whose label sets are obtained through certain set operations on the label sets of the
corresponding input pairs.
思想自由 兼容并包
More of an
expectation.
Takes content into
consideration
Training process is different. We take
intersection of the labels (albeit in
feature space) and not the content.
What would happen if intersection of
labels if empty set.
思想自由 兼容并包
Methodology
𝒁𝒊𝒏𝒕 = 𝑴𝒊𝒏𝒕(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒖𝒏𝒊 = 𝑴 𝒖𝒏𝒊(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒔𝒖𝒃 = 𝑴 𝒔𝒖𝒃(𝑭 𝒙, 𝑭 𝒚)
思想自由 兼容并包
Binary Cross-Entropy Multi-Label Loss -
思想自由 兼容并包
Methodology
𝐶𝑙𝑜𝑠𝑠
𝒁𝒊𝒏𝒕 = 𝑴𝒊𝒏𝒕(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒖𝒏𝒊 = 𝑴 𝒖𝒏𝒊(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒔𝒖𝒃 = 𝑴 𝒔𝒖𝒃(𝑭 𝒙, 𝑭 𝒚)
思想自由 兼容并包
Methodology
𝐿𝑎𝑠𝑜𝑙𝑜𝑠𝑠
𝒁𝒊𝒏𝒕 = 𝑴𝒊𝒏𝒕(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒖𝒏𝒊 = 𝑴 𝒖𝒏𝒊(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒔𝒖𝒃 = 𝑴 𝒔𝒖𝒃(𝑭 𝒙, 𝑭 𝒚)
MLP with 3 or 4 blocks
思想自由 兼容并包
Methodology
𝑅𝑙𝑜𝑠𝑠
𝑠𝑦𝑚
𝒁𝒊𝒏𝒕 = 𝑴𝒊𝒏𝒕(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒖𝒏𝒊 = 𝑴 𝒖𝒏𝒊(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒔𝒖𝒃 = 𝑴 𝒔𝒖𝒃(𝑭 𝒙, 𝑭 𝒚)
思想自由 兼容并包
Methodology
𝒁𝒊𝒏𝒕 = 𝑴𝒊𝒏𝒕(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒖𝒏𝒊 = 𝑴 𝒖𝒏𝒊(𝑭 𝒙, 𝑭 𝒚)
𝒁 𝒔𝒖𝒃 = 𝑴 𝒔𝒖𝒃(𝑭 𝒙, 𝑭 𝒚)
𝑅𝑙𝑜𝑠𝑠
𝑚𝑐 mc – mode
collapse
思想自由 兼容并包
思想自由 兼容并包
Experiments
Coco: object-based multi-label dataset (80 object categories)
Celeb-A: attribute-based multi-label dataset(40 facial attributes)
思想自由 兼容并包
Coco Experiments: Classification
The Coco dataset is split into Coco-A(64 classes), Coco-B(16 classes)
in order to evaluate the model on the unseen
categories, they separately train a 16-way classifier
on them, used for evaluation. This is a standard
metric in domain generalization, but it means the
model can not confuse them with already seen
classes, which would be a likely source of errors
思想自由 兼容并包
Coco Experiments: Retrieval
The Coco dataset is split into Coco-A(64 classes), Coco-B(16 classes)
Their backbone itself seems to be
okish. Would like to see its
performance numbers (guess based
on visual results)
Why IOU now, we
are only bothered
about labels, no?
思想自由 兼容并包
Experiments
Celeb-A: attribute-based multi-label dataset
(40 facial attributes)
(No seen unseen in this scenario)
思想自由 兼容并包
Experiments: Ablations
Coco: object-based multi-label dataset
Celeb-A: attribute-based multi-label dataset (40 facial attributes)
Tab. Analytic approximations to set operations
All numbers are mAP %
Inspired by DCGAN
思想自由 兼容并包
Visual Results
思想自由 兼容并包
Source - https://ameroyer.github.io/reading-notes/few-shot%20learning/2019/04/26/laso_label_set_operations_networks_for_multi_label_few_shot_learning.html

More Related Content

Similar to LaSo

Martin Odersky: What's next for Scala
Martin Odersky: What's next for ScalaMartin Odersky: What's next for Scala
Martin Odersky: What's next for ScalaMarakana Inc.
 
ShawnQuinnCSS581FinalProjectReport
ShawnQuinnCSS581FinalProjectReportShawnQuinnCSS581FinalProjectReport
ShawnQuinnCSS581FinalProjectReportShawn Quinn
 
StackNet Meta-Modelling framework
StackNet Meta-Modelling frameworkStackNet Meta-Modelling framework
StackNet Meta-Modelling frameworkSri Ambati
 
DotNetFest - Let’s refresh our memory! Memory management in .NET
DotNetFest - Let’s refresh our memory! Memory management in .NETDotNetFest - Let’s refresh our memory! Memory management in .NET
DotNetFest - Let’s refresh our memory! Memory management in .NETMaarten Balliauw
 
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management....NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...NETFest
 
Kernel based swarm optimization for renewable energy application
Kernel based swarm optimization  for renewable energy applicationKernel based swarm optimization  for renewable energy application
Kernel based swarm optimization for renewable energy applicationAboul Ella Hassanien
 
ScalaDays 2013 Keynote Speech by Martin Odersky
ScalaDays 2013 Keynote Speech by Martin OderskyScalaDays 2013 Keynote Speech by Martin Odersky
ScalaDays 2013 Keynote Speech by Martin OderskyTypesafe
 
Exploring .NET memory management - JetBrains webinar
Exploring .NET memory management - JetBrains webinarExploring .NET memory management - JetBrains webinar
Exploring .NET memory management - JetBrains webinarMaarten Balliauw
 
Scala overview
Scala overviewScala overview
Scala overviewSteve Min
 
Towards Computational Research Objects
Towards Computational Research ObjectsTowards Computational Research Objects
Towards Computational Research ObjectsDavid De Roure
 
06 InheritanceAndPolymorphism.ppt
06 InheritanceAndPolymorphism.ppt06 InheritanceAndPolymorphism.ppt
06 InheritanceAndPolymorphism.pptParikhitGhosh1
 
Google guava overview
Google guava overviewGoogle guava overview
Google guava overviewSteve Min
 
Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...
Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...
Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...Politecnico di Milano
 
Shuvrojit Majumder . 25900120006 Object Oriented Programming (PCC-CS 503) ...
Shuvrojit Majumder .  25900120006  Object Oriented Programming (PCC-CS 503)  ...Shuvrojit Majumder .  25900120006  Object Oriented Programming (PCC-CS 503)  ...
Shuvrojit Majumder . 25900120006 Object Oriented Programming (PCC-CS 503) ...ShuvrojitMajumder
 
Php meetup 20130912 reflection
Php meetup 20130912   reflectionPhp meetup 20130912   reflection
Php meetup 20130912 reflectionJoelRSimpson
 
ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in...
 ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in... ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in...
ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in...Saurabh Nanda
 
Experiments with Pro-Active Declarative Meta-Programming
Experiments with Pro-Active Declarative Meta-ProgrammingExperiments with Pro-Active Declarative Meta-Programming
Experiments with Pro-Active Declarative Meta-ProgrammingESUG
 
Object oriented javascript
Object oriented javascriptObject oriented javascript
Object oriented javascriptUsman Mehmood
 
The Anatomy of Analysis Tools (EVO 2008)
The Anatomy of Analysis Tools (EVO 2008)The Anatomy of Analysis Tools (EVO 2008)
The Anatomy of Analysis Tools (EVO 2008)Tudor Girba
 

Similar to LaSo (20)

Martin Odersky: What's next for Scala
Martin Odersky: What's next for ScalaMartin Odersky: What's next for Scala
Martin Odersky: What's next for Scala
 
ShawnQuinnCSS581FinalProjectReport
ShawnQuinnCSS581FinalProjectReportShawnQuinnCSS581FinalProjectReport
ShawnQuinnCSS581FinalProjectReport
 
StackNet Meta-Modelling framework
StackNet Meta-Modelling frameworkStackNet Meta-Modelling framework
StackNet Meta-Modelling framework
 
DotNetFest - Let’s refresh our memory! Memory management in .NET
DotNetFest - Let’s refresh our memory! Memory management in .NETDotNetFest - Let’s refresh our memory! Memory management in .NET
DotNetFest - Let’s refresh our memory! Memory management in .NET
 
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management....NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...
 
Kernel based swarm optimization for renewable energy application
Kernel based swarm optimization  for renewable energy applicationKernel based swarm optimization  for renewable energy application
Kernel based swarm optimization for renewable energy application
 
ScalaDays 2013 Keynote Speech by Martin Odersky
ScalaDays 2013 Keynote Speech by Martin OderskyScalaDays 2013 Keynote Speech by Martin Odersky
ScalaDays 2013 Keynote Speech by Martin Odersky
 
Exploring .NET memory management - JetBrains webinar
Exploring .NET memory management - JetBrains webinarExploring .NET memory management - JetBrains webinar
Exploring .NET memory management - JetBrains webinar
 
Java Basics
Java BasicsJava Basics
Java Basics
 
Scala overview
Scala overviewScala overview
Scala overview
 
Towards Computational Research Objects
Towards Computational Research ObjectsTowards Computational Research Objects
Towards Computational Research Objects
 
06 InheritanceAndPolymorphism.ppt
06 InheritanceAndPolymorphism.ppt06 InheritanceAndPolymorphism.ppt
06 InheritanceAndPolymorphism.ppt
 
Google guava overview
Google guava overviewGoogle guava overview
Google guava overview
 
Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...
Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...
Enriching Intelligent Textbooks with Interactivity: When Smart Content Alloca...
 
Shuvrojit Majumder . 25900120006 Object Oriented Programming (PCC-CS 503) ...
Shuvrojit Majumder .  25900120006  Object Oriented Programming (PCC-CS 503)  ...Shuvrojit Majumder .  25900120006  Object Oriented Programming (PCC-CS 503)  ...
Shuvrojit Majumder . 25900120006 Object Oriented Programming (PCC-CS 503) ...
 
Php meetup 20130912 reflection
Php meetup 20130912   reflectionPhp meetup 20130912   reflection
Php meetup 20130912 reflection
 
ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in...
 ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in... ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in...
ABRIDGED VERSION - Joys & frustrations of putting 34,000 lines of Haskell in...
 
Experiments with Pro-Active Declarative Meta-Programming
Experiments with Pro-Active Declarative Meta-ProgrammingExperiments with Pro-Active Declarative Meta-Programming
Experiments with Pro-Active Declarative Meta-Programming
 
Object oriented javascript
Object oriented javascriptObject oriented javascript
Object oriented javascript
 
The Anatomy of Analysis Tools (EVO 2008)
The Anatomy of Analysis Tools (EVO 2008)The Anatomy of Analysis Tools (EVO 2008)
The Anatomy of Analysis Tools (EVO 2008)
 

Recently uploaded

CARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptxCARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptxGaneshChakor2
 
Crayon Activity Handout For the Crayon A
Crayon Activity Handout For the Crayon ACrayon Activity Handout For the Crayon A
Crayon Activity Handout For the Crayon AUnboundStockton
 
BASLIQ CURRENT LOOKBOOK LOOKBOOK(1) (1).pdf
BASLIQ CURRENT LOOKBOOK  LOOKBOOK(1) (1).pdfBASLIQ CURRENT LOOKBOOK  LOOKBOOK(1) (1).pdf
BASLIQ CURRENT LOOKBOOK LOOKBOOK(1) (1).pdfSoniaTolstoy
 
Hybridoma Technology ( Production , Purification , and Application )
Hybridoma Technology  ( Production , Purification , and Application  ) Hybridoma Technology  ( Production , Purification , and Application  )
Hybridoma Technology ( Production , Purification , and Application ) Sakshi Ghasle
 
Presiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha electionsPresiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha electionsanshu789521
 
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPTECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPTiammrhaywood
 
Introduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher EducationIntroduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher Educationpboyjonauth
 
Accessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impactAccessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impactdawncurless
 
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17Celine George
 
Software Engineering Methodologies (overview)
Software Engineering Methodologies (overview)Software Engineering Methodologies (overview)
Software Engineering Methodologies (overview)eniolaolutunde
 
Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111Sapana Sha
 
Mastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory InspectionMastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory InspectionSafetyChain Software
 
Paris 2024 Olympic Geographies - an activity
Paris 2024 Olympic Geographies - an activityParis 2024 Olympic Geographies - an activity
Paris 2024 Olympic Geographies - an activityGeoBlogs
 
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptxContemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptxRoyAbrique
 
Introduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptxIntroduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptxpboyjonauth
 
Sanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdfSanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdfsanyamsingh5019
 
18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdf18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdfssuser54595a
 
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...Krashi Coaching
 

Recently uploaded (20)

CARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptxCARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptx
 
Crayon Activity Handout For the Crayon A
Crayon Activity Handout For the Crayon ACrayon Activity Handout For the Crayon A
Crayon Activity Handout For the Crayon A
 
BASLIQ CURRENT LOOKBOOK LOOKBOOK(1) (1).pdf
BASLIQ CURRENT LOOKBOOK  LOOKBOOK(1) (1).pdfBASLIQ CURRENT LOOKBOOK  LOOKBOOK(1) (1).pdf
BASLIQ CURRENT LOOKBOOK LOOKBOOK(1) (1).pdf
 
Hybridoma Technology ( Production , Purification , and Application )
Hybridoma Technology  ( Production , Purification , and Application  ) Hybridoma Technology  ( Production , Purification , and Application  )
Hybridoma Technology ( Production , Purification , and Application )
 
Presiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha electionsPresiding Officer Training module 2024 lok sabha elections
Presiding Officer Training module 2024 lok sabha elections
 
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPTECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
ECONOMIC CONTEXT - LONG FORM TV DRAMA - PPT
 
9953330565 Low Rate Call Girls In Rohini Delhi NCR
9953330565 Low Rate Call Girls In Rohini  Delhi NCR9953330565 Low Rate Call Girls In Rohini  Delhi NCR
9953330565 Low Rate Call Girls In Rohini Delhi NCR
 
Introduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher EducationIntroduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher Education
 
Accessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impactAccessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impact
 
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
Incoming and Outgoing Shipments in 1 STEP Using Odoo 17
 
Software Engineering Methodologies (overview)
Software Engineering Methodologies (overview)Software Engineering Methodologies (overview)
Software Engineering Methodologies (overview)
 
Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111Call Girls in Dwarka Mor Delhi Contact Us 9654467111
Call Girls in Dwarka Mor Delhi Contact Us 9654467111
 
Mastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory InspectionMastering the Unannounced Regulatory Inspection
Mastering the Unannounced Regulatory Inspection
 
Paris 2024 Olympic Geographies - an activity
Paris 2024 Olympic Geographies - an activityParis 2024 Olympic Geographies - an activity
Paris 2024 Olympic Geographies - an activity
 
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptxContemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
 
Introduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptxIntroduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptx
 
Sanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdfSanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdf
 
18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdf18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAСY_INDEX-DM_23-1-final-eng.pdf
 
Staff of Color (SOC) Retention Efforts DDSD
Staff of Color (SOC) Retention Efforts DDSDStaff of Color (SOC) Retention Efforts DDSD
Staff of Color (SOC) Retention Efforts DDSD
 
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
Kisan Call Centre - To harness potential of ICT in Agriculture by answer farm...
 

LaSo

Editor's Notes

  1. Fewshot: Only a handful of examples are available for the task