Search for: All records

Language models are optimized to learn which responses you prefer, but they don't learn why you preferred a particular response. This limits their ability to tailor to personalized requests (e.g., "What should I eat for dinner? I'm vegetarian"), so we introduce a simple fix: have models infer personas that explain why users could prefer responses. We show training on these inferred personas leads to responses that are significantly more personalized for user needs. 

Language models like ChatGPT are pretty good at answering questions (e.g. "What is 12 * 12?"), but we show they can surprisingly struggle when asked to do the reverse task: generating questions for answers (e.g. "Give me a question with the answer 144"). We study when these errors happen, what might be causing them, and how they can be addressed. 

Abstract With rapid progress in simulation of strongly interacting quantum Hamiltonians, the challenge in characterizing unknown phases becomes a bottleneck for scientific progress. We demonstrate that a Quantum-Classical hybrid approach (QuCl) of mining sampled projective snapshots with interpretable classical machine learning can unveil signatures of seemingly featureless quantum states. The Kitaev-Heisenberg model on a honeycomb lattice under external magnetic field presents an ideal system to test QuCl, where simulations have found an intermediate gapless phase (IGP) sandwiched between known phases, launching a debate over its elusive nature. We use the correlator convolutional neural network, trained on labeled projective snapshots, in conjunction with regularization path analysis to identify signatures of phases. We show that QuCl reproduces known features of established phases. Significantly, we also identify a signature of the IGP in the spin channel perpendicular to the field direction, which we interpret as a signature of Friedel oscillations of gapless spinons forming a Fermi surface. Our predictions can guide future experimental searches for spin liquids. 

Peer prediction refers to a collection of mechanisms for eliciting information from human agents when direct verification of the obtained information is unavailable. They are designed to have a game-theoretic equilibrium where everyone reveals their private information truthfully. This result holds under the assumption that agents are Bayesian and they each adopt a fixed strategy across all tasks. Human agents however are observed in many domains to exhibit learning behavior in sequential settings. In this paper, we explore the dynamics of sequential peer prediction mechanisms when participants are learning agents. We first show that the notion of no regret alone for the agents’ learning algorithms cannot guaran- tee convergence to the truthful strategy. We then focus on a family of learning algorithms where strategy updates only depend on agents’ cumulative rewards and prove that agents’ strategies in the popular Correlated Agreement (CA) mechanism converge to truthful reporting when they use algorithms from this family. This fam- ily of algorithms is not necessarily no-regret, but includes several familiar no-regret learning algorithms (e.g multiplicative weight update and Follow the Perturbed Leader) as special cases. Simulation of several algorithms in this family as well as the ε-greedy algorithm, which is outside of this family, shows convergence to the truthful strategy in the CA mechanism. 

Graph Neural Networks (GNNs) have emerged as a powerful tool for semi- supervised node classification tasks. However, recent studies have revealed various biases in GNNs stemming from both node features and graph topology. In this work, we uncover a new bias - label position bias, which indicates that the node closer to the labeled nodes tends to perform better. We introduce a new metric, the Label Proximity Score, to quantify this bias, and find that it is closely related to performance disparities. To address the label position bias, we propose a novel optimization framework for learning a label position unbiased graph structure, which can be applied to existing GNNs. Extensive experiments demonstrate that our proposed method not only outperforms backbone methods but also significantly mitigates the issue of label position bias in GNNs. 

Abstract The glutaminase enzymes GAC and GLS2 catalyze the hydrolysis of glutamine to glutamate, satisfying the ‘glutamine addiction’ of cancer cells. They are the targets of anti-cancer drugs; however, their mechanisms of activation and catalytic activity have been unclear. Here we demonstrate that the ability of GAC and GLS2 to form filaments is directly coupled to their catalytic activity and present their cryo-EM structures which provide a view of the conformational states essential for catalysis. Filament formation guides an ‘activation loop’ to assume a specific conformation that works together with a ‘lid’ to close over the active site and position glutamine for nucleophilic attack by an essential serine. Our findings highlight how ankyrin repeats on GLS2 regulate enzymatic activity, while allosteric activators stabilize, and clinically relevant inhibitors block, filament formation that enables glutaminases to catalyze glutaminolysis and support cancer progression. 

Learning vocabulary (e.g., benevolent) can be tedious, but using mnemonics (e.g., benevolent sounds like "benefits," and a kind boss gives benefits) makes it more engaging and effective. This paper introduces SMART, a large language model trained to produce mnemonics based on feedback from flashcard learners. Students struggle to predict which mnemonics will help them most. Still, by training SMART on both student preferences and learning outcomes, we can generate mnemonics as effectively as GPT-4, but at a much lower cost.