Proposes safety and reliability weighted detection metrics using a target criticality model; evaluated nine detectors on nuScenes.

Andrea Ceccarelli, Leonardo Montecchi

Fine-tuning GPT-3 with human preferences and PPO reinforcement learning improves instruction-following, truthfulness, and safety; parameters reduced from 175B to 1.3B.

Long Ouyang, Jeff Wu, Xu Jiang et al.

SyFES framework automatically constructs symbolic density functionals; GAS22 outperforms on MGCDB84 test set.

He Ma, Arunachalam Narayanaswamy, Patrick Riley et al.

ACCEL method evolves complexity by editing levels, significantly enhancing RL agent capabilities.

Jack Parker-Holder, Minqi Jiang, Michael Dennis et al.

GFlowNets combined with epistemic uncertainty enable diverse, high-quality biological sequence generation for active learning.

Moksh Jain, Emmanuel Bengio, Alex-Hernandez Garcia et al.

DeepNorm stabilizes extremely deep Transformers, enabling training up to 1000 layers with significant performance gains in multilingual translation.

Hongyu Wang, Shuming Ma, Li Dong et al.

Proposes a theory of abstraction with three key criteria to improve sample efficiency and generalization in reinforcement learning.

David Abel

Introduces weighted quantile conformal prediction for nonexchangeable data, improving robustness under distribution shifts.

Rina Foygel Barber, Emmanuel J. Candes, Aaditya Ramdas et al.

Proposed a trainable subgraph retriever (SR) that enhances multi-hop KBQA by decoupling retrieval and reasoning, achieving state-of-the-art results.

Jing Zhang, Xiaokang Zhang, Jifan Yu et al.

This study reveals that in in-context learning, models rely more on demonstration format and distribution than on correct labels.

Sewon Min, Xinxi Lyu, Ari Holtzman et al.

BagPipe system accelerates deep recommendation model training by up to 5.6× via caching and lookahead prefetching.

Saurabh Agarwal, Chengpo Yan, Ziyi Zhang et al.

Proposes an average-case error bound for kernel-based Bayesian quadrature using a two-step meta-algorithm, effective in noisy and randomized settings, especially with Matérn and SE kernels.

Xu Cai, Chi Thanh Lam, Jonathan Scarlett

Proposes SSCD, a self-supervised contrastive descriptor for image copy detection, outperforming baselines by 48% on DISC2021.

Ed Pizzi, Sreya Dutta Roy, Sugosh Nagavara Ravindra et al.

This study reveals that fine-tuning can distort pretrained features, harming out-of-distribution generalization; proposes LP-FT strategy for balanced performance.

Ananya Kumar, Aditi Raghunathan, Robbie Jones et al.

Proposes the merged-staircase property (MSP) as a necessary and nearly sufficient condition for SGD learning of sparse functions on two-layer neural networks.

Emmanuel Abbe, Enric Boix-Adsera, Theodor Misiakiewicz

Proposes forward gradient method using forward automatic differentiation to estimate gradients without backpropagation, enabling training speeds up to twice as fast.

Atılım Güneş Baydin, Barak A. Pearlmutter, Don Syme et al.

Proposes a continuous conditional GAN (CC-GAN) for 3D point cloud generation, achieving explicit size control with low regression error (0.28%) and superior quality (FPD 1.5290).

Larissa T. Triess, Andre Bühler, David Peter et al.

Ditto reconstructs articulated digital twins from before/after point clouds, reaching 0.72 whole-object Chamfer Distance on Shape2Motion.

Zhenyu Jiang, Cheng-Chun Hsu, Yuke Zhu

A computer control method using reinforcement learning and behavioral cloning achieves human-level performance on the MiniWob++ benchmark.

Peter C Humphreys, David Raposo, Toby Pohlen et al.

Study reveals predictability and surprise in large generative models, exploring social impact.

Deep Ganguli, Danny Hernandez, Liane Lovitt et al.