Open problems in technical ai governance
AI progress is creating a growing range of risks and opportunities, but it is often unclear how
they should be navigated. In many cases, the barriers and uncertainties faced are at least …
they should be navigated. In many cases, the barriers and uncertainties faced are at least …
Zero-knowledge proofs of training for deep neural networks
A zero-knowledge proof of training (zkPoT) enables a party to prove that they have correctly
trained a committed model based on a committed dataset without revealing any additional …
trained a committed model based on a committed dataset without revealing any additional …
Causes and effects of unanticipated numerical deviations in neural network inference frameworks
Hardware-specific optimizations in machine learning (ML) frameworks can cause numerical
deviations of inference results. Quite surprisingly, despite using a fixed trained model and …
deviations of inference results. Quite surprisingly, despite using a fixed trained model and …
Unforgeability in stochastic gradient descent
Stochastic Gradient Descent (SGD) is a popular training algorithm, a cornerstone of modern
machine learning systems. Several security applications benefit from determining if SGD …
machine learning systems. Several security applications benefit from determining if SGD …
Proof-of-learning with incentive security
Most concurrent blockchain systems rely heavily on the Proof-of-Work (PoW) or Proof-of-
Stake (PoS) mechanisms for decentralized consensus and security assurance. However, the …
Stake (PoS) mechanisms for decentralized consensus and security assurance. However, the …
Gradients look alike: Sensitivity is often overestimated in {DP-SGD}
Differentially private stochastic gradient descent (DP-SGD) is the canonical approach to
private deep learning. While the current privacy analysis of DP-SGD is known to be tight in …
private deep learning. While the current privacy analysis of DP-SGD is known to be tight in …
Survey on Blockchain-Enhanced Machine Learning
The convergence of blockchain and Machine Learning (ML) promises to reshape
technological innovation by enhancing security, efficiency, and transparency in ML systems …
technological innovation by enhancing security, efficiency, and transparency in ML systems …
Attesting distributional properties of training data for machine learning
The success of machine learning (ML) has been accompanied by increased concerns about
its trustworthiness. Several jurisdictions are preparing ML regulatory frameworks. One such …
its trustworthiness. Several jurisdictions are preparing ML regulatory frameworks. One such …
Zkdl: Efficient zero-knowledge proofs of deep learning training
The recent advancements in deep learning have brought about significant changes in
various aspects of people's lives. Meanwhile, these rapid developments have raised …
various aspects of people's lives. Meanwhile, these rapid developments have raised …
Enhancing Security of Proof-of-Learning against Spoofing Attacks using Feature-Based Model Watermarking
The rapid advancement of machine learning (ML) technologies necessitates robust security
frameworks to protect the integrity of ML model training processes. Proof-of-Learning (PoL) …
frameworks to protect the integrity of ML model training processes. Proof-of-Learning (PoL) …