SlideShare a Scribd company logo
1 of 13
Download to read offline
Learning Prototype Classifiers for Long-Tailed Recognition
Saurabh
Sharma1
1 University of California Santa Barbara, USA
Ning
Yu3
Yongqin
Xian2
Ambuj
Singh1
2 Google, Switzerland
3 Salesforce Research, USA
{saurabhsharma,ambuj}@cs.ucsb.edu
yxian@google.com ning.yu@salesforce.com
IJCAI 23, Macao
Imbalanced distributions in real world datasets
2
Distribution of training images per species. iNat2017
Applications- Autonomous
driving, object detection,
fraud detection, eliminating
bias in ML models
IJCAI 23, Macao
Long-Tailed Recognition
• Problem formulation: Given a long-tailed training set,
maximize accuracy on a balanced test set.
• Prior work:
‣ Loss reshaping: Focal loss, Class-balanced loss, LDAM loss, Logit adjustment.
‣ Ensembles: Class-balanced experts, LFME, BBN, RIDE.
‣ Others: Decoupled training, Weight decay regularization, data augmentation, self-
supervised pre-training
3
Key challenges:
1. Relative imbalance
2. Data scarcity
IJCAI 23, Macao
LTR Using Biased Linear Softmax
• Linear softmax classi
fi
ers have both a direction and a magnitude.
• The direction closely aligns with the class means (neural collapse).
• However, the magnitude gets correlated to the label distribution prior
, leading to biased decision boundaries.
μy
p(y)
4
IJCAI 23, Macao
Prototype Classi
fi
ers for LTR
• We propose distance-based classi
fi
cation using learnable
prototypes.
• Prototype classi
fi
ers outperform linear softmax and
nearest-class-mean classi
fi
ers.
• Our theoretical analysis shows that prototype classi
fi
ers
overcome the biased softmax problem.
5
IJCAI 23, Macao
Learning Prototype Classi
fi
ers
• We compute pre-softmax logit scores using distances:
where are
fi
xed representations from a baseline model,
and are learnable class prototypes.
• Inference is done using the nearest-prototype rule:
log p(y|g(x)) = −
1
2
d(g(x), cy)
g(x)
cy
6
IJCAI 23, Macao
Choice of distance metric
• Euclidean distance:
• Stable gradient updates on prototypes:
• L2 norm of gradient is independent of .
• Only depends on the probability of mis-classi
fi
cation.
• Optimization is robust to outliers that have a high .
d(g(x), cy)
d(g(x), cy)
7
IJCAI 23, Macao
Addressing the biased softmax problem
• We show that the prototype classi
fi
er is a linear softmax classi
fi
er,
where:
• Bias term negates the gains from increasing or decreasing
the norm of the weight term.
• The prototype classi
fi
er is robust to imbalanced distributions.
8
weight: cy
bias: −
∥cy∥2
2
IJCAI 23, Macao
Channel-dependent temperatures
• As distance scales vary along each channel, we use
channel-dependent temperatures:
• High T Low sensitivity
Low T High sensitivity
• Generalized Mahalanobis distance metric.
⟹
⟹
9
IJCAI 23, Macao
Prototype Classi
fi
er learns equi-norm prototypes
10
IJCAI 23, Macao
Learnt prototypes are well-separated
11
Average Euclidean distance Average cosine similarity
IJCAI 23, Macao 12
CIFAR 100-LT
ImageNet-LT
iNaturalist18
Comparison to the
state-of-the-arts
IJCAI 23, Macao
Conclusion
• We present Learnable Prototype Classi
fi
ers for LTR.
• Prototype Classi
fi
ers overcome the intrinsic bias of linear softmax
classi
fi
ers and are robust to imbalanced distributions.
• Euclidean distance based prototype classi
fi
ers are robust to outliers
because of its stable gradient property.
• Learnt prototypes are equi-norm and well-separated.
• For more details, please take a look below:
13
Code
Paper

More Related Content

Similar to Learning Prototype Classifiers for Long-Tailed Recognition

Generalized Linear Models in Spark MLlib and SparkR
Generalized Linear Models in Spark MLlib and SparkRGeneralized Linear Models in Spark MLlib and SparkR
Generalized Linear Models in Spark MLlib and SparkRDatabricks
 
DutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in MLDutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in MLBigML, Inc
 
PR-305: Exploring Simple Siamese Representation Learning
PR-305: Exploring Simple Siamese Representation LearningPR-305: Exploring Simple Siamese Representation Learning
PR-305: Exploring Simple Siamese Representation LearningSungchul Kim
 
generalized_nbody_acs_2015_challacombe
generalized_nbody_acs_2015_challacombegeneralized_nbody_acs_2015_challacombe
generalized_nbody_acs_2015_challacombeMatt Challacombe
 
PR-393: ResLT: Residual Learning for Long-tailed Recognition
PR-393: ResLT: Residual Learning for Long-tailed RecognitionPR-393: ResLT: Residual Learning for Long-tailed Recognition
PR-393: ResLT: Residual Learning for Long-tailed RecognitionSunghoon Joo
 
Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...Dalei Li
 
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...adil raja
 
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...adil raja
 
ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...
ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...
ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...Jihun Yun
 
Classification using L1-Penalized Logistic Regression
Classification using L1-Penalized Logistic RegressionClassification using L1-Penalized Logistic Regression
Classification using L1-Penalized Logistic RegressionSetia Pramana
 
ngboost.pptx
ngboost.pptxngboost.pptx
ngboost.pptxHadrian7
 
CAMCOS_final Presentation_Group2.pptx
CAMCOS_final Presentation_Group2.pptxCAMCOS_final Presentation_Group2.pptx
CAMCOS_final Presentation_Group2.pptxhuzaifaazam3
 
Human uncertainty makes classification more robust, ICCV 2019 Review
Human uncertainty makes classification more robust, ICCV 2019 ReviewHuman uncertainty makes classification more robust, ICCV 2019 Review
Human uncertainty makes classification more robust, ICCV 2019 ReviewLEE HOSEONG
 
Study on Application of Ensemble learning on Credit Scoring
Study on Application of Ensemble learning on Credit ScoringStudy on Application of Ensemble learning on Credit Scoring
Study on Application of Ensemble learning on Credit Scoringharmonylab
 

Similar to Learning Prototype Classifiers for Long-Tailed Recognition (20)

Generalized Linear Models in Spark MLlib and SparkR
Generalized Linear Models in Spark MLlib and SparkRGeneralized Linear Models in Spark MLlib and SparkR
Generalized Linear Models in Spark MLlib and SparkR
 
Kulakov_Mykhailo.pptx
Kulakov_Mykhailo.pptxKulakov_Mykhailo.pptx
Kulakov_Mykhailo.pptx
 
DutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in MLDutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in ML
 
ARIMA
ARIMA ARIMA
ARIMA
 
PR-305: Exploring Simple Siamese Representation Learning
PR-305: Exploring Simple Siamese Representation LearningPR-305: Exploring Simple Siamese Representation Learning
PR-305: Exploring Simple Siamese Representation Learning
 
generalized_nbody_acs_2015_challacombe
generalized_nbody_acs_2015_challacombegeneralized_nbody_acs_2015_challacombe
generalized_nbody_acs_2015_challacombe
 
PR-393: ResLT: Residual Learning for Long-tailed Recognition
PR-393: ResLT: Residual Learning for Long-tailed RecognitionPR-393: ResLT: Residual Learning for Long-tailed Recognition
PR-393: ResLT: Residual Learning for Long-tailed Recognition
 
Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...Two strategies for large-scale multi-label classification on the YouTube-8M d...
Two strategies for large-scale multi-label classification on the YouTube-8M d...
 
Support Vector Machines
Support Vector MachinesSupport Vector Machines
Support Vector Machines
 
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
 
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
Modeling the Effect of Packet Loss on Speech Quality: Genetic Programming Bas...
 
230727_HB_JointJournalClub.pptx
230727_HB_JointJournalClub.pptx230727_HB_JointJournalClub.pptx
230727_HB_JointJournalClub.pptx
 
ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...
ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...
ProxGen: Adaptive Proximal Gradient Methods for Structured Neural Networks (N...
 
EE5180_G-5.pptx
EE5180_G-5.pptxEE5180_G-5.pptx
EE5180_G-5.pptx
 
Classification using L1-Penalized Logistic Regression
Classification using L1-Penalized Logistic RegressionClassification using L1-Penalized Logistic Regression
Classification using L1-Penalized Logistic Regression
 
ngboost.pptx
ngboost.pptxngboost.pptx
ngboost.pptx
 
MUMS: Transition & SPUQ Workshop - Some Strategies to Quantify Uncertainty fo...
MUMS: Transition & SPUQ Workshop - Some Strategies to Quantify Uncertainty fo...MUMS: Transition & SPUQ Workshop - Some Strategies to Quantify Uncertainty fo...
MUMS: Transition & SPUQ Workshop - Some Strategies to Quantify Uncertainty fo...
 
CAMCOS_final Presentation_Group2.pptx
CAMCOS_final Presentation_Group2.pptxCAMCOS_final Presentation_Group2.pptx
CAMCOS_final Presentation_Group2.pptx
 
Human uncertainty makes classification more robust, ICCV 2019 Review
Human uncertainty makes classification more robust, ICCV 2019 ReviewHuman uncertainty makes classification more robust, ICCV 2019 Review
Human uncertainty makes classification more robust, ICCV 2019 Review
 
Study on Application of Ensemble learning on Credit Scoring
Study on Application of Ensemble learning on Credit ScoringStudy on Application of Ensemble learning on Credit Scoring
Study on Application of Ensemble learning on Credit Scoring
 

Recently uploaded

Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...Sérgio Sacani
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisDiwakar Mishra
 
Artificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C PArtificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C PPRINCE C P
 
Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.aasikanpl
 
Natural Polymer Based Nanomaterials
Natural Polymer Based NanomaterialsNatural Polymer Based Nanomaterials
Natural Polymer Based NanomaterialsAArockiyaNisha
 
G9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.pptG9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.pptMAESTRELLAMesa2
 
Biopesticide (2).pptx .This slides helps to know the different types of biop...
Biopesticide (2).pptx  .This slides helps to know the different types of biop...Biopesticide (2).pptx  .This slides helps to know the different types of biop...
Biopesticide (2).pptx .This slides helps to know the different types of biop...RohitNehra6
 
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptxUnlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptxanandsmhk
 
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCESTERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCEPRINCE C P
 
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...Sérgio Sacani
 
A relative description on Sonoporation.pdf
A relative description on Sonoporation.pdfA relative description on Sonoporation.pdf
A relative description on Sonoporation.pdfnehabiju2046
 
Biological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfBiological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfmuntazimhurra
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSarthak Sekhar Mondal
 
Types of different blotting techniques.pptx
Types of different blotting techniques.pptxTypes of different blotting techniques.pptx
Types of different blotting techniques.pptxkhadijarafiq2012
 
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxAnalytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxSwapnil Therkar
 
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxSOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxkessiyaTpeter
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxgindu3009
 
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...jana861314
 

Recently uploaded (20)

Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
 
Artificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C PArtificial Intelligence In Microbiology by Dr. Prince C P
Artificial Intelligence In Microbiology by Dr. Prince C P
 
Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
Call Girls in Mayapuri Delhi 💯Call Us 🔝9953322196🔝 💯Escort.
 
Natural Polymer Based Nanomaterials
Natural Polymer Based NanomaterialsNatural Polymer Based Nanomaterials
Natural Polymer Based Nanomaterials
 
G9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.pptG9 Science Q4- Week 1-2 Projectile Motion.ppt
G9 Science Q4- Week 1-2 Projectile Motion.ppt
 
Biopesticide (2).pptx .This slides helps to know the different types of biop...
Biopesticide (2).pptx  .This slides helps to know the different types of biop...Biopesticide (2).pptx  .This slides helps to know the different types of biop...
Biopesticide (2).pptx .This slides helps to know the different types of biop...
 
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptxUnlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptx
 
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCESTERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
STERILITY TESTING OF PHARMACEUTICALS ppt by DR.C.P.PRINCE
 
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
 
A relative description on Sonoporation.pdf
A relative description on Sonoporation.pdfA relative description on Sonoporation.pdf
A relative description on Sonoporation.pdf
 
Biological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfBiological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdf
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
 
Types of different blotting techniques.pptx
Types of different blotting techniques.pptxTypes of different blotting techniques.pptx
Types of different blotting techniques.pptx
 
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxAnalytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
 
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxSOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
 
CELL -Structural and Functional unit of life.pdf
CELL -Structural and Functional unit of life.pdfCELL -Structural and Functional unit of life.pdf
CELL -Structural and Functional unit of life.pdf
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptx
 
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
 
9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service
9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service
9953056974 Young Call Girls In Mahavir enclave Indian Quality Escort service
 

Learning Prototype Classifiers for Long-Tailed Recognition

  • 1. Learning Prototype Classifiers for Long-Tailed Recognition Saurabh Sharma1 1 University of California Santa Barbara, USA Ning Yu3 Yongqin Xian2 Ambuj Singh1 2 Google, Switzerland 3 Salesforce Research, USA {saurabhsharma,ambuj}@cs.ucsb.edu yxian@google.com ning.yu@salesforce.com
  • 2. IJCAI 23, Macao Imbalanced distributions in real world datasets 2 Distribution of training images per species. iNat2017 Applications- Autonomous driving, object detection, fraud detection, eliminating bias in ML models
  • 3. IJCAI 23, Macao Long-Tailed Recognition • Problem formulation: Given a long-tailed training set, maximize accuracy on a balanced test set. • Prior work: ‣ Loss reshaping: Focal loss, Class-balanced loss, LDAM loss, Logit adjustment. ‣ Ensembles: Class-balanced experts, LFME, BBN, RIDE. ‣ Others: Decoupled training, Weight decay regularization, data augmentation, self- supervised pre-training 3 Key challenges: 1. Relative imbalance 2. Data scarcity
  • 4. IJCAI 23, Macao LTR Using Biased Linear Softmax • Linear softmax classi fi ers have both a direction and a magnitude. • The direction closely aligns with the class means (neural collapse). • However, the magnitude gets correlated to the label distribution prior , leading to biased decision boundaries. μy p(y) 4
  • 5. IJCAI 23, Macao Prototype Classi fi ers for LTR • We propose distance-based classi fi cation using learnable prototypes. • Prototype classi fi ers outperform linear softmax and nearest-class-mean classi fi ers. • Our theoretical analysis shows that prototype classi fi ers overcome the biased softmax problem. 5
  • 6. IJCAI 23, Macao Learning Prototype Classi fi ers • We compute pre-softmax logit scores using distances: where are fi xed representations from a baseline model, and are learnable class prototypes. • Inference is done using the nearest-prototype rule: log p(y|g(x)) = − 1 2 d(g(x), cy) g(x) cy 6
  • 7. IJCAI 23, Macao Choice of distance metric • Euclidean distance: • Stable gradient updates on prototypes: • L2 norm of gradient is independent of . • Only depends on the probability of mis-classi fi cation. • Optimization is robust to outliers that have a high . d(g(x), cy) d(g(x), cy) 7
  • 8. IJCAI 23, Macao Addressing the biased softmax problem • We show that the prototype classi fi er is a linear softmax classi fi er, where: • Bias term negates the gains from increasing or decreasing the norm of the weight term. • The prototype classi fi er is robust to imbalanced distributions. 8 weight: cy bias: − ∥cy∥2 2
  • 9. IJCAI 23, Macao Channel-dependent temperatures • As distance scales vary along each channel, we use channel-dependent temperatures: • High T Low sensitivity Low T High sensitivity • Generalized Mahalanobis distance metric. ⟹ ⟹ 9
  • 10. IJCAI 23, Macao Prototype Classi fi er learns equi-norm prototypes 10
  • 11. IJCAI 23, Macao Learnt prototypes are well-separated 11 Average Euclidean distance Average cosine similarity
  • 12. IJCAI 23, Macao 12 CIFAR 100-LT ImageNet-LT iNaturalist18 Comparison to the state-of-the-arts
  • 13. IJCAI 23, Macao Conclusion • We present Learnable Prototype Classi fi ers for LTR. • Prototype Classi fi ers overcome the intrinsic bias of linear softmax classi fi ers and are robust to imbalanced distributions. • Euclidean distance based prototype classi fi ers are robust to outliers because of its stable gradient property. • Learnt prototypes are equi-norm and well-separated. • For more details, please take a look below: 13 Code Paper