Picture for Baharan Mirzasoleiman

Baharan Mirzasoleiman

Make the Most of Your Data: Changing the Training Data Distribution to Improve In-distribution Generalization Performance

Add code
Apr 27, 2024
Viaarxiv icon

Data-Efficient Contrastive Language-Image Pretraining: Prioritizing Data Quality over Quantity

Add code
Mar 20, 2024
Figure 1 for Data-Efficient Contrastive Language-Image Pretraining: Prioritizing Data Quality over Quantity
Figure 2 for Data-Efficient Contrastive Language-Image Pretraining: Prioritizing Data Quality over Quantity
Figure 3 for Data-Efficient Contrastive Language-Image Pretraining: Prioritizing Data Quality over Quantity
Figure 4 for Data-Efficient Contrastive Language-Image Pretraining: Prioritizing Data Quality over Quantity
Viaarxiv icon

Investigating the Benefits of Projection Head for Representation Learning

Add code
Mar 18, 2024
Figure 1 for Investigating the Benefits of Projection Head for Representation Learning
Figure 2 for Investigating the Benefits of Projection Head for Representation Learning
Figure 3 for Investigating the Benefits of Projection Head for Representation Learning
Figure 4 for Investigating the Benefits of Projection Head for Representation Learning
Viaarxiv icon

SmallToLarge (S2L): Scalable Data Selection for Fine-tuning Large Language Models by Summarizing Training Trajectories of Small Models

Add code
Mar 12, 2024
Figure 1 for SmallToLarge (S2L): Scalable Data Selection for Fine-tuning Large Language Models by Summarizing Training Trajectories of Small Models
Figure 2 for SmallToLarge (S2L): Scalable Data Selection for Fine-tuning Large Language Models by Summarizing Training Trajectories of Small Models
Figure 3 for SmallToLarge (S2L): Scalable Data Selection for Fine-tuning Large Language Models by Summarizing Training Trajectories of Small Models
Figure 4 for SmallToLarge (S2L): Scalable Data Selection for Fine-tuning Large Language Models by Summarizing Training Trajectories of Small Models
Viaarxiv icon

Inference and Interference: The Role of Clipping, Pruning and Loss Landscapes in Differentially Private Stochastic Gradient Descent

Add code
Nov 12, 2023
Viaarxiv icon

Data Distillation Can Be Like Vodka: Distilling More Times For Better Quality

Add code
Oct 10, 2023
Figure 1 for Data Distillation Can Be Like Vodka: Distilling More Times For Better Quality
Figure 2 for Data Distillation Can Be Like Vodka: Distilling More Times For Better Quality
Figure 3 for Data Distillation Can Be Like Vodka: Distilling More Times For Better Quality
Figure 4 for Data Distillation Can Be Like Vodka: Distilling More Times For Better Quality
Viaarxiv icon

Understanding the Robustness of Multi-modal Contrastive Learning to Distribution Shift

Add code
Oct 08, 2023
Figure 1 for Understanding the Robustness of Multi-modal Contrastive Learning to Distribution Shift
Figure 2 for Understanding the Robustness of Multi-modal Contrastive Learning to Distribution Shift
Figure 3 for Understanding the Robustness of Multi-modal Contrastive Learning to Distribution Shift
Figure 4 for Understanding the Robustness of Multi-modal Contrastive Learning to Distribution Shift
Viaarxiv icon

Better Safe than Sorry: Pre-training CLIP against Targeted Data Poisoning and Backdoor Attacks

Add code
Oct 05, 2023
Figure 1 for Better Safe than Sorry: Pre-training CLIP against Targeted Data Poisoning and Backdoor Attacks
Figure 2 for Better Safe than Sorry: Pre-training CLIP against Targeted Data Poisoning and Backdoor Attacks
Figure 3 for Better Safe than Sorry: Pre-training CLIP against Targeted Data Poisoning and Backdoor Attacks
Figure 4 for Better Safe than Sorry: Pre-training CLIP against Targeted Data Poisoning and Backdoor Attacks
Viaarxiv icon

Ordering for Non-Replacement SGD

Add code
Jun 28, 2023
Figure 1 for Ordering for Non-Replacement SGD
Figure 2 for Ordering for Non-Replacement SGD
Figure 3 for Ordering for Non-Replacement SGD
Figure 4 for Ordering for Non-Replacement SGD
Viaarxiv icon

Towards Mitigating Spurious Correlations in the Wild: A Benchmark & a more Realistic Dataset

Add code
Jun 21, 2023
Figure 1 for Towards Mitigating Spurious Correlations in the Wild: A Benchmark & a more Realistic Dataset
Figure 2 for Towards Mitigating Spurious Correlations in the Wild: A Benchmark & a more Realistic Dataset
Figure 3 for Towards Mitigating Spurious Correlations in the Wild: A Benchmark & a more Realistic Dataset
Figure 4 for Towards Mitigating Spurious Correlations in the Wild: A Benchmark & a more Realistic Dataset
Viaarxiv icon