Picture for Shang-Tse Chen

Shang-Tse Chen

Academia Sinica

SYN2REAL: Leveraging Task Arithmetic for Mitigating Synthetic-Real Discrepancies in ASR Domain Adaptation

Add code
Jun 05, 2024
Viaarxiv icon

Step by Step to Fairness: Attributing Societal Bias in Task-oriented Dialogue Systems

Add code
Nov 14, 2023
Figure 1 for Step by Step to Fairness: Attributing Societal Bias in Task-oriented Dialogue Systems
Figure 2 for Step by Step to Fairness: Attributing Societal Bias in Task-oriented Dialogue Systems
Figure 3 for Step by Step to Fairness: Attributing Societal Bias in Task-oriented Dialogue Systems
Figure 4 for Step by Step to Fairness: Attributing Societal Bias in Task-oriented Dialogue Systems
Viaarxiv icon

Learning from Red Teaming: Gender Bias Provocation and Mitigation in Large Language Models

Add code
Oct 17, 2023
Viaarxiv icon

Annealing Self-Distillation Rectification Improves Adversarial Training

Add code
May 20, 2023
Figure 1 for Annealing Self-Distillation Rectification Improves Adversarial Training
Figure 2 for Annealing Self-Distillation Rectification Improves Adversarial Training
Figure 3 for Annealing Self-Distillation Rectification Improves Adversarial Training
Figure 4 for Annealing Self-Distillation Rectification Improves Adversarial Training
Viaarxiv icon

Position Matters! Empirical Study of Order Effect in Knowledge-grounded Dialogue

Add code
Feb 12, 2023
Figure 1 for Position Matters! Empirical Study of Order Effect in Knowledge-grounded Dialogue
Figure 2 for Position Matters! Empirical Study of Order Effect in Knowledge-grounded Dialogue
Figure 3 for Position Matters! Empirical Study of Order Effect in Knowledge-grounded Dialogue
Figure 4 for Position Matters! Empirical Study of Order Effect in Knowledge-grounded Dialogue
Viaarxiv icon

QCRS: Improve Randomized Smoothing using Quasi-Concave Optimization

Add code
Feb 01, 2023
Figure 1 for QCRS: Improve Randomized Smoothing using Quasi-Concave Optimization
Figure 2 for QCRS: Improve Randomized Smoothing using Quasi-Concave Optimization
Figure 3 for QCRS: Improve Randomized Smoothing using Quasi-Concave Optimization
Figure 4 for QCRS: Improve Randomized Smoothing using Quasi-Concave Optimization
Viaarxiv icon

Fair Robust Active Learning by Joint Inconsistency

Add code
Sep 22, 2022
Figure 1 for Fair Robust Active Learning by Joint Inconsistency
Figure 2 for Fair Robust Active Learning by Joint Inconsistency
Figure 3 for Fair Robust Active Learning by Joint Inconsistency
Figure 4 for Fair Robust Active Learning by Joint Inconsistency
Viaarxiv icon

Enhancing Targeted Attack Transferability via Diversified Weight Pruning

Add code
Aug 18, 2022
Figure 1 for Enhancing Targeted Attack Transferability via Diversified Weight Pruning
Figure 2 for Enhancing Targeted Attack Transferability via Diversified Weight Pruning
Figure 3 for Enhancing Targeted Attack Transferability via Diversified Weight Pruning
Figure 4 for Enhancing Targeted Attack Transferability via Diversified Weight Pruning
Viaarxiv icon

Few-shot Prompting Towards Controllable Response Generation

Add code
Jun 09, 2022
Figure 1 for Few-shot Prompting Towards Controllable Response Generation
Figure 2 for Few-shot Prompting Towards Controllable Response Generation
Figure 3 for Few-shot Prompting Towards Controllable Response Generation
Figure 4 for Few-shot Prompting Towards Controllable Response Generation
Viaarxiv icon

UnMask: Adversarial Detection and Defense Through Robust Feature Alignment

Add code
Feb 21, 2020
Figure 1 for UnMask: Adversarial Detection and Defense Through Robust Feature Alignment
Figure 2 for UnMask: Adversarial Detection and Defense Through Robust Feature Alignment
Figure 3 for UnMask: Adversarial Detection and Defense Through Robust Feature Alignment
Figure 4 for UnMask: Adversarial Detection and Defense Through Robust Feature Alignment
Viaarxiv icon