Micah Goldblum

发表

Just How Toxic is Data Poisoning? A Unified Benchmark for Backdoor and Data Poisoning Attacks pdf

Micah Goldblum, Tom Goldstein, Avi Schwarzschild, 2020, ICML.

Adversarially Robust Few-Shot Learning: A Meta-Learning Approach

T. Goldstein, Micah Goldblum, Liam Fowl, 2019, NeurIPS.

Data Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses

A. Madry, Xinyun Chen, D. Song, 2020 .

Adversarial Attacks on Machine Learning Systems for High-Frequency Trading

Ankit B. Patel, T. Goldstein, Avi Schwarzschild, 2020, ArXiv.

LowKey: Leveraging Adversarial Attacks to Protect Social Media Users from Facial Recognition pdf

Micah Goldblum, Tom Goldstein, Gavin Taylor, 2021, ICLR.

The Intrinsic Dimension of Images and Its Impact on Learning pdf

Micah Goldblum, Tom Goldstein, Chen Zhu, 2021, ICLR.

Analyzing the Machine Learning Conference Review Process pdf

Micah Goldblum, Tom Goldstein, David Tran, 2020, ArXiv.

Can You Learn an Algorithm? Generalizing from Easy to Hard Problems with Recurrent Networks pdf

Uzi Vishkin, Micah Goldblum, Tom Goldstein, 2021, NeurIPS.

A Closer Look at Distribution Shifts and Out-of-Distribution Generalization on Graphs

Furong Huang, D. Wipf, T. Goldstein, 2021 .

Transfer Learning with Deep Tabular Models

A. G. Wilson, T. Goldstein, Avi Schwarzschild, 2022, ICLR.

Truth or Backpropaganda? An Empirical Investigation of Deep Learning Theory

T. Goldstein, Avi Schwarzschild, Micah Goldblum, 2019, ICLR.

Autoregressive Perturbations for Data Poisoning

T. Goldstein, Micah Goldblum, Vasu Singla, 2022, ArXiv.

The Close Relationship Between Contrastive Learning and Meta-Learning

T. Goldstein, Micah Goldblum, Renkun Ni, 2022, ICLR.

Sleeper Agent: Scalable Hidden Trigger Backdoors for Neural Networks Trained from Scratch

Ramalingam Chellappa, T. Goldstein, Micah Goldblum, 2021, NeurIPS.

Where do Models go Wrong? Parameter-Space Saliency Maps for Explainability

Furong Huang, T. Goldstein, Micah Goldblum, 2021, NeurIPS.

Canary in a Coalmine: Better Membership Inference with Ensembled Adversarial Queries

T. Goldstein, Micah Goldblum, Jonas Geiping, 2022, ArXiv.

Autoregressive Perturbations for Data Poisoning

T. Goldstein, Micah Goldblum, Vasu Singla, 2022, NeurIPS.

Fishing for User Data in Large-Batch Federated Learning via Gradient Magnification

T. Goldstein, Micah Goldblum, Liam Fowl, 2022, ICML.

How Much Data Are Augmentations Worth? An Investigation into Scaling Laws, Invariance, and Implicit Regularization

T. Goldstein, Micah Goldblum, Ravid Shwartz-Ziv, 2022, ArXiv.

End-to-end Algorithm Synthesis with Recurrent Networks: Logical Extrapolation Without Overthinking

Furong Huang, T. Goldstein, Avi Schwarzschild, 2022, ArXiv.

Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models

T. Goldstein, Micah Goldblum, Vasu Singla, 2022, 2212.03860.

Cold Diffusion: Inverting Arbitrary Image Transforms Without Noise

Furong Huang, T. Goldstein, Micah Goldblum, 2022, NeurIPS.

What do Vision Transformers Learn? A Visual Exploration

Micah Goldblum, T. Goldstein, Manli Shu, 2022, arXiv.org.

K-SAM: Sharpness-Aware Minimization at the Speed of SGD

T. Goldstein, Micah Goldblum, Ping-Yeh Chiang, 2022, ArXiv.

A Deep Dive into Dataset Imbalance and Bias in Face Identification

T. Goldstein, Micah Goldblum, Valeriia Cherepanova, 2022, AIES.

Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models

T. Goldstein, Micah Goldblum, Vasu Singla, 2022, 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses

A. Madry, Xinyun Chen, D. Song, 2020, IEEE Transactions on Pattern Analysis and Machine Intelligence.

SAINT: Improved Neural Networks for Tabular Data via Row Attention and Contrastive Pre-Training pdf

Micah Goldblum, Tom Goldstein, Avi Schwarzschild, 2021, ArXiv.

Strong Data Augmentation Sanitizes Poisoning and Backdoor Attacks Without an Accuracy Tradeoff pdf

Jonas Geiping, Micah Goldblum, Liam Fowl, 2020, ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

How Much Data Are Augmentations Worth? An Investigation into Scaling Laws, Invariance, and Implicit Regularization

T. Goldstein, Micah Goldblum, Ravid Shwartz-Ziv, 2022, ICLR.

A Cookbook of Self-Supervised Learning

Ari S. Morcos, Yann LeCun, H. Pirsiavash, 2023, ArXiv.

Understanding Generalization through Visualizations pdf

Micah Goldblum, Liam Fowl, Tom Goldstein, 2019, ICBINB@NeurIPS.

Can Neural Nets Learn the Same Model Twice? Investigating Reproducibility and Double Descent from the Decision Boundary Perspective

Richard Baraniuk, T. Goldstein, Micah Goldblum, 2022, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Technical Challenges for Training Fair Neural Networks pdf

Micah Goldblum, Tom Goldstein, John P. Dickerson, 2021, ArXiv.

Data Augmentation for Meta-Learning

Micah Goldblum, Tom Goldstein, Amr Sharaf, 2020, ICML.

Thinking Two Moves Ahead: Anticipating Other Users Improves Backdoor Attacks in Federated Learning

Ramalingam Chellappa, T. Goldstein, Micah Goldblum, 2022, ArXiv.

Decepticons: Corrupted Transformers Breach Privacy in Federated Learning for Language Models

T. Goldstein, Micah Goldblum, Liam Fowl, 2022, ArXiv.

Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery

T. Goldstein, Micah Goldblum, Jonas Geiping, 2023, arXiv.org.

Universal Guidance for Diffusion Models

T. Goldstein, Avi Schwarzschild, Micah Goldblum, 2023, 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

Exploring and Exploiting Decision Boundary Dynamics for Adversarial Robustness

Furong Huang, T. Goldstein, Micah Goldblum, 2023, ICLR.

Just How Toxic is Data Poisoning? A Unified Benchmark for Backdoor and Data Poisoning Attacks

John P. Dickerson, T. Goldstein, Avi Schwarzschild, 2020, ICML.

Unraveling Meta-Learning: Understanding Feature Representations for Few-Shot Tasks

T. Goldstein, Micah Goldblum, Renkun Ni, 2020, ICML.

Adversarial Examples Make Strong Poisons pdf

Jonas Geiping, Micah Goldblum, Liam Fowl, 2021, NeurIPS.

Adversarial attacks on machine learning systems for high-frequency trading pdf

Micah Goldblum, Tom Goldstein, Avi Schwarzschild, 2020, ICAIF.

Adversarially Robust Few-Shot Learning: A Meta-Learning Approach

Micah Goldblum, Liam Fowl, Tom Goldstein, 2019, NeurIPS.

Adversarially Robust Distillation

Micah Goldblum, Liam Fowl, Tom Goldstein, 2019, AAAI.

MetaBalance: High-Performance Neural Networks for Class-Imbalanced Data pdf

Micah Goldblum, Tom Goldstein, Avi Schwarzschild, 2021, ArXiv.

Poisons that are learned faster are more effective

T. Goldstein, Micah Goldblum, Liam Fowl, 2022, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

DP-InstaHide: Provably Defusing Poisoning and Backdoor Attacks with Differentially Private Data Augmentations

Furong Huang, T. Goldstein, Micah Goldblum, 2021, ArXiv.

What Doesn't Kill You Makes You Robust(er): How to Adversarially Train against Data Poisoning

T. Goldstein, Micah Goldblum, Liam Fowl, 2021, 2102.13624.

Preventing Unauthorized Use of Proprietary Data: Poisoning for Secure Dataset Release pdf

Jonas Geiping, Micah Goldblum, Liam Fowl, 2021, ArXiv.

What Doesn't Kill You Makes You Robust(er): Adversarial Training against Poisons and Backdoors

T. Goldstein, Micah Goldblum, Liam Fowl, 2021, ArXiv.

Canary in a Coalmine: Better Membership Inference with Ensembled Adversarial Queries

T. Goldstein, Micah Goldblum, Jonas Geiping, 2022, ArXiv.

Witchcraft: Efficient PGD Attacks with Random Step Size

Jonas Geiping, Micah Goldblum, Tom Goldstein, 2019, ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery

T. Goldstein, Micah Goldblum, Jonas Geiping, 2023, ArXiv.

Plug-In Inversion: Model-Agnostic Inversion for Vision with Data Augmentations

T. Goldstein, Micah Goldblum, Chen Zhu, 2022, ICML.

A Cookbook of Self-Supervised Learning

Ari S. Morcos, Yann LeCun, H. Pirsiavash, 2023, ArXiv.

Understanding and Mitigating Copying in Diffusion Models

T. Goldstein, Micah Goldblum, Vasu Singla, 2023, ArXiv.

Bring Your Own Data! Self-Supervised Evaluation for Large Language Models

T. Goldstein, Micah Goldblum, Jonas Geiping, 2023, ArXiv.

On the Reliability of Watermarks for Large Language Models

T. Goldstein, Micah Goldblum, Jonas Geiping, 2023, ArXiv.

VESTIGATION OF DEEP LEARNING THEORY

Micah Goldblum, Jonas Geiping, 2020 .

The Uncanny Similarity of Recurrence and Depth

T. Goldstein, Avi Schwarzschild, Micah Goldblum, 2021, ICLR.

Transfer Learning with Deep Tabular Models

T. Goldstein, Avi Schwarzschild, Micah Goldblum, 2022, ICLR.

Baseline Defenses for Adversarial Attacks Against Aligned Language Models

Avi Schwarzschild, Micah Goldblum, Jonas Geiping, 2023, ArXiv.

What Can We Learn from Unlearnable Datasets?

T. Goldstein, Micah Goldblum, Vasu Singla, 2023, ArXiv.

A Performance-Driven Benchmark for Feature Selection in Tabular Deep Learning

Micah Goldblum, Jonas Geiping, Valeriia Cherepanova, 2023, NeurIPS.

Battle of the Backbones: A Large-Scale Comparison of Pretrained Models across Computer Vision Tasks

Ramalingam Chellappa, Micah Goldblum, Renkun Ni, 2023, NeurIPS.

A Simple and Efficient Baseline for Data Attribution on Images

Micah Goldblum, Vasu Singla, Jonas Geiping, 2023, ArXiv.

NEFTune: Noisy Embeddings Improve Instruction Finetuning

Avi Schwarzschild, Micah Goldblum, B. Kailkhura, 2023, ArXiv.

Decepticons: Corrupted Transformers Breach Privacy in Federated Learning for Language Models

T. Goldstein, Micah Goldblum, Jonas Geiping, 2022, ICLR.

Panning for Gold in Federated Learning: Targeted Text Extraction under Arbitrarily Large-Scale Aggregation

T. Goldstein, Micah Goldblum, Jonas Geiping, 2023, ICLR.

Perspectives on the State and Future of Deep Learning - 2023

Zachary Chase Lipton, Micah Goldblum, Preetum Nakkiran, 2023, ArXiv.

Towards Transferable Adversarial Attacks on Image and Video Transformers

Micah Goldblum, Zuxuan Wu, Zhipeng Wei, 2023, IEEE Transactions on Image Processing.

LiveBench: A Challenging, Contamination-Free LLM Benchmark

W. Neiswanger, Micah Goldblum, Ravid Shwartz-Ziv, 2024, ArXiv.