Seong Min Kye*, Kwanghee Choi*, Hyeongmin Byun, Buru Chang�Hyperconnect Inc. (*Same contribution)
TiDAL 🌊 : Learning Training Dynamics�for Active Learning
CONTENT
Active Learning
Typical Settings of Active Learning
1. Preliminaries
Active Learning Methods
Measuring the data uncertainty
1. Preliminaries
Long-tailed Classification�����
2. Pilot Study
Model Snapshot vs. Training Dynamics
p(T): Model prediction� on the final (T-th) epoch�
p̅(T): Averaged model predictions� during training (T epochs)
2. Pilot Study
Theorem 1
Under the LE-SDE framework, with the assumption of local elasticity,�certain samples and uncertain samples reveal different TD; especially, certain samples converge quickly than uncertain samples.
Theorem 2
Estimators such as Entropy and Margin successfully capture the difference of TD between easy and hard samples even for the case where it cannot be distinguished via the predicted probabilities of the model snapshot.
3. Theoretical Results
4. Our Method
5. Results and Analyses
Datasets
Baseline Methods
Balanced Datasets
5. Results and Analyses
Imbalanced Datasets
5. Results and Analyses
5. Results and Analyses
Ablation Study
Performance of the TD prediction module
Seong Min Kye harris@hpcnt.com
Kwanghee Choi kwanghec@andrew.cmu.edu
Hyeongmin Byun boris@hpcnt.com
Buru Chang buru@sogang.ac.kr
Thank you for listening!
TiDAL: Learning Training Dynamics for Active Learning
Links
Contacts