Search Results - "Pfau, Jacob"

Stress testing reveals gaps in clinic readiness of image-based diagnostic artificial intelligence models by Young, Albert T., Fernandez, Kristen, Pfau, Jacob, Reddy, Rasika, Cao, Nhat Anh, von Franque, Max Y., Johal, Arjun, Wu, Benjamin V., Wu, Rachel R., Chen, Jennifer Y., Fadadu, Raj P., Vasquez, Juan A., Tam, Andrew, Keiser, Michael J., Wei, Maria L.

“…Artificial intelligence models match or exceed dermatologists in melanoma image classification. Less is known about their robustness against real-world…”
Get full text

Artificial Intelligence in Dermatology: A Primer by Young, Albert T., Xiong, Mulin, Pfau, Jacob, Keiser, Michael J., Wei, Maria L.

“…Artificial intelligence is becoming increasingly important in dermatology, with studies reporting accuracy matching or exceeding dermatologists for the…”
Get full text

Artificial Intelligence in Teledermatology by Xiong, Mulin, Pfau, Jacob, Young, Albert T., Wei, Maria L.

“…Purpose of Review This review summarizes current and prospective applications of artificial intelligence (AI) and smartphone technologies to automated…”
Get full text

Let's Think Dot by Dot: Hidden Computation in Transformer Language Models by Pfau, Jacob, Merrill, William, Bowman, Samuel R

“…Chain-of-thought responses from language models improve performance across most benchmarks. However, it remains unclear to what extent these performance gains…”
Get full text

Steering Without Side Effects: Improving Post-Deployment Control of Language Models by Stickland, Asa Cooper, Lyzhov, Alexander, Pfau, Jacob, Mahdi, Salsabila, Bowman, Samuel R

“…Language models (LMs) have been shown to behave unexpectedly post-deployment. For example, new jailbreaks continually arise, allowing model misuse, despite…”
Get full text

Taking AI Welfare Seriously by Long, Robert, Sebo, Jeff, Butlin, Patrick, Finlinson, Kathleen, Fish, Kyle, Harding, Jacqueline, Pfau, Jacob, Sims, Toni, Birch, Jonathan, Chalmers, David

“…In this report, we argue that there is a realistic possibility that some AI systems will be conscious and/or robustly agentic in the near future. That means…”
Get full text

Self-Consistency of Large Language Models under Ambiguity by Bartsch, Henning, Jorgensen, Ole, Rosati, Domenic, Hoelscher-Obermaier, Jason, Pfau, Jacob

“…Large language models (LLMs) that do not give consistent answers across contexts are problematic when used for tasks with expectations of consistency, e.g.,…”
Get full text

Goal Misgeneralization in Deep Reinforcement Learning by Langosco, Lauro, Koch, Jack, Sharkey, Lee, Pfau, Jacob, Orseau, Laurent, Krueger, David

“…We study goal misgeneralization, a type of out-of-distribution generalization failure in reinforcement learning (RL). Goal misgeneralization failures occur…”
Get full text

Robust Semantic Interpretability: Revisiting Concept Activation Vectors by Pfau, Jacob, Young, Albert T, Wei, Jerome, Wei, Maria L, Keiser, Michael J

“…Interpretability methods for image classification assess model trustworthiness by attempting to expose whether the model is systematically biased or attending…”
Get full text

Global Saliency: Aggregating Saliency Maps to Assess Dataset Artefact Bias by Pfau, Jacob, Young, Albert T, Wei, Maria L, Keiser, Michael J

“…In high-stakes applications of machine learning models, interpretability methods provide guarantees that models are right for the right reasons. In medical…”
Get full text

Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback by Casper, Stephen, Davies, Xander, Shi, Claudia, Gilbert, Thomas Krendl, Scheurer, Jérémy, Rando, Javier, Freedman, Rachel, Korbak, Tomasz, Lindner, David, Freire, Pedro, Wang, Tony, Marks, Samuel, Segerie, Charbel-Raphaël, Carroll, Micah, Peng, Andi, Christoffersen, Phillip, Damani, Mehul, Slocum, Stewart, Anwar, Usman, Siththaranjan, Anand, Nadeau, Max, Michaud, Eric J, Pfau, Jacob, Krasheninnikov, Dmitrii, Chen, Xin, Langosco, Lauro, Hase, Peter, Bıyık, Erdem, Dragan, Anca, Krueger, David, Sadigh, Dorsa, Hadfield-Menell, Dylan

“…Reinforcement learning from human feedback (RLHF) is a technique for training AI systems to align with human goals. RLHF has emerged as the central method used…”
Get full text