159 citations · 509 across the 29 of their papers we have counts for
31 papers
Do Theorem Provers Need Theorem-Proving Agents? mini-Lean-agent: A Coding-Agent Baseline for Lean
Tomas Ortega, Simon Park, Ayush Khaitan +3
General-purpose coding agents use standard shell and file interactions, yet formal theorem-proving agents often rely on specialized, ad-hoc tooling. This makes it difficult to dete…
Rethinking On-Policy Self-Distillation for Thinking Models
Simran Kaur, Narutatsu Ri, Yinghui He +2
Self-distillation is a promising recipe for self-improvement in language models. In this setting, a model can serve as its own teacher when given privileged information, such as a…
Goedel-Architect: Streamlining Formal Theorem Proving with Blueprint Generation and Refinement
Jui-Hui Chung, Ziyang Cai, Zihao Li +14
We introduce Goedel-Architect, an agentic framework for formal theorem proving in Lean 4 centered on blueprint generation and refinement. A blueprint is a dependency graph of defin…
To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning
Nevena Lazić, Liam Fowl, András György +1
We investigate the ability of decoder-only transformer models to perform abstract symbolic reasoning; specifically solving propositional logic reasoning problems given in-context.…
Self-Distillation Zero: Self-Revision Turns Binary Rewards into Dense Supervision
Yinghui He, Simran Kaur, Adithya Bhaskar +7
Current post-training methods in verifiable settings fall into two categories. Reinforcement learning (RLVR) relies on binary rewards, which are broadly applicable and powerful, bu…
Escaping the Cognitive Well: Efficient Competition Math with Off-the-Shelf Models
Xingyu Dang, Rohit Agarwal, Rodrigo Porto +3
In the past year, custom and unreleased math reasoning models reached gold medal performance on the International Mathematical Olympiad (IMO). Similar performance was then reported…
3D Gaussian Splatting with Normal Information for Mesh Extraction and Improved Rendering
Meenakshi Krishnan, Liam Fowl, Ramani Duraiswami
Differentiable 3D Gaussian splatting has emerged as an efficient and flexible rendering technique for representing complex scenes from a collection of 2D views and enabling high-qu…
Generating Potent Poisons and Backdoors from Scratch with Guided Diffusion
Hossein Souri, Arpit Bansal, Hamid Kazemi +7
Modern neural networks are often trained on massive datasets that are web scraped with minimal human inspection. As a result of this insecure curation pipeline, an adversary can po…
Exploring Sequence-to-Sequence Transformer-Transducer Models for Keyword Spotting
Beltrán Labrador, Guanlong Zhao, Ignacio López Moreno +3
In this paper, we present a novel approach to adapt a sequence-to-sequence Transformer-Transducer ASR system to the keyword spotting (KWS) task. We achieve this by replacing the ke…
Thinking Two Moves Ahead: Anticipating Other Users Improves Backdoor Attacks in Federated Learning
Yuxin Wen, Jonas Geiping, Liam Fowl +4
Federated learning is particularly susceptible to model poisoning and backdoor attacks because individual users have direct control over the training data and model updates. At the…
Poisons that are learned faster are more effective
Pedro Sandoval-Segura, Vasu Singla, Liam Fowl +4
Imperceptible poisoning attacks on entire datasets have recently been touted as methods for protecting data privacy. However, among a number of defenses preventing the practical us…
Can Neural Nets Learn the Same Model Twice? Investigating Reproducibility and Double Descent from the Decision Boundary Perspective
Gowthami Somepalli, Liam Fowl, Arpit Bansal +5
We discuss methods for visualizing neural network decision boundaries and decision regions. We use these visualizations to investigate issues related to reproducibility and general…
Fishing for User Data in Large-Batch Federated Learning via Gradient Magnification
Yuxin Wen, Jonas Geiping, Liam Fowl +2
Federated learning (FL) has rapidly risen in popularity due to its promise of privacy and efficiency. Previous works have exposed privacy vulnerabilities in the FL pipeline by reco…
Decepticons: Corrupted Transformers Breach Privacy in Federated Learning for Language Models
Liam Fowl, Jonas Geiping, Steven Reich +4
A central tenet of Federated learning (FL), which trains models without centralizing user data, is privacy. However, previous work has shown that the gradient updates used in FL ca…
Execute Order 66: Targeted Data Poisoning for Reinforcement Learning
Harrison Foley, Liam Fowl, Tom Goldstein +1
Data poisoning for reinforcement learning has historically focused on general performance degradation, and targeted attacks have been successful via perturbations that involve cont…
Robbing the Fed: Directly Obtaining Private Data in Federated Learning with Modified Models
Liam Fowl, Jonas Geiping, Wojtek Czaja +2
Federated learning has quickly gained popularity with its promises of increased user privacy and efficiency. Previous works have shown that federated gradient updates contain infor…
Adversarial Examples Make Strong Poisons
Liam Fowl, Micah Goldblum, Ping-yeh Chiang +3
The adversarial machine learning literature is largely partitioned into evasion attacks on testing data and poisoning attacks on training data. In this work, we show that adversari…
Sleeper Agent: Scalable Hidden Trigger Backdoors for Neural Networks Trained from Scratch
Hossein Souri, Liam Fowl, Rama Chellappa +2
As the curation of data for machine learning becomes increasingly automated, dataset tampering is a mounting threat. Backdoor attackers tamper with training data to embed a vulnera…
DP-InstaHide: Provably Defusing Poisoning and Backdoor Attacks with Differentially Private Data Augmentations
Eitan Borgnia, Jonas Geiping, Valeriia Cherepanova +6
Data poisoning and backdoor attacks manipulate training data to induce security breaches in a victim model. These attacks can be provably deflected using differentially private (DP…
Preventing Unauthorized Use of Proprietary Data: Poisoning for Secure Dataset Release
Liam Fowl, Ping-yeh Chiang, Micah Goldblum +4
Large organizations such as social media companies continually release data, for example user images. At the same time, these organizations leverage their massive corpora of releas…
What Doesn't Kill You Makes You Robust(er): How to Adversarially Train against Data Poisoning
Jonas Geiping, Liam Fowl, Gowthami Somepalli +3
Data poisoning is a threat model in which a malicious actor tampers with training data to manipulate outcomes at inference time. A variety of defenses against this threat model hav…
Strong Data Augmentation Sanitizes Poisoning and Backdoor Attacks Without an Accuracy Tradeoff
Eitan Borgnia, Valeriia Cherepanova, Liam Fowl +5
Data poisoning and backdoor attacks manipulate victim models by maliciously modifying training data. In light of this growing threat, a recent survey of industry professionals reve…
Random Network Distillation as a Diversity Metric for Both Image and Text Generation
Liam Fowl, Micah Goldblum, Arjun Gupta +2
Generative models are increasingly able to produce remarkably high quality images and text. The community has developed numerous evaluation metrics for comparing generative models.…
Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching
Jonas Geiping, Liam Fowl, W. Ronny Huang +4
Data Poisoning attacks modify training data to maliciously control a model trained on such data. In this work, we focus on targeted poisoning attacks which cause a reclassification…