Proposes 3DILG with irregular latent grids, boosting 3D shape reconstruction and generation performance.

Biao Zhang, Matthias Nießner, Peter Wonka

Quark algorithm uses quantized reward conditioning to generate text with reduced undesirable traits, improving quality.

Ximing Lu, Sean Welleck, Jack Hessel et al.

Mathematical analysis reveals traditional RPE in Transformers limits universal approximation; proposes URPE to ensure model expressiveness.

Shengjie Luo, Shanda Li, Shuxin Zheng et al.

OptFormer, a Transformer-based universal hyperparameter optimizer, can imitate 7 algorithms, leveraging large-scale real tuning data to improve prediction and optimization.

Yutian Chen, Xingyou Song, Chansoo Lee et al.

Introduces a Monte Carlo tree search-based symbolic physics learner (SPL) that automatically discovers nonlinear differential equations from limited noisy data, outperforming state-of-the-art methods.

Fangzheng Sun, Yang Liu, Jian-Xun Wang et al.

Proposes Contextual Pandora’s Box with reservation value linking context and distribution, achieving sublinear regret.

Alexia Atsidakou, Constantine Caramanis, Evangelia Gergatsouli et al.

NaturalProver leverages background references with constrained decoding to generate and suggest mathematical proofs, achieving over 40% correctness and usefulness in human evaluations.

Sean Welleck, Jiacheng Liu, Ximing Lu et al.

Proposes a simple estimator for total treatment effect in randomized experiments without network structure knowledge, with statistical guarantees.

Christina Lee Yu, Edoardo M Airoldi, Christian Borgs et al.

Using large language models for autoformalization, achieving 25.3% perfect translation of math problems into Isabelle/HOL.

Yuhuai Wu, Albert Q. Jiang, Wenda Li et al.

Neural demographic perturbation enhances NLP fairness by reducing bias sensitivity, validated on GLUE and bias datasets.

Rebecca Qian, Candace Ross, Jude Fernandes et al.

This study shows pre-trained models inherently regularize, making KD and label smoothing less effective during fine-tuning; performance differences are negligible.

Ivan Kobyzev, Aref Jafari, Mehdi Rezagholizadeh et al.

Multi-Head Online Learning model significantly improves delayed feedback CVR and VPC predictions.

Hui Gao, Yihan Yang

Proposes Zero-shot-CoT, using simple prompts to boost large LLMs' zero-shot reasoning, improving MultiArith accuracy from 17.7% to 78.7%.

Takeshi Kojima, Shixiang Shane Gu, Machel Reid et al.

This paper introduces ALTI+, a layer-wise token contribution method for Transformer-based encoder-decoder models, quantifying source and target influences on translation outputs.

Javier Ferrando, Gerard I. Gállego, Belen Alastruey et al.

Imagen combines large-scale frozen T5-XXL encoder with diffusion models, achieving a state-of-the-art FID of 7.27 on COCO, greatly enhancing photorealism.

Chitwan Saharia, William Chan, Saurabh Saxena et al.

StreamingQA uses time-stamped news data to evaluate model adaptation, showing fine-tuning and retrieval updates improve performance in dynamic knowledge environments.

Adam Liška, Tomáš Kočiský, Elena Gribovskaya et al.

BBTv2 optimizes large models' prompts using a gradient-free algorithm, reducing parameters with performance comparable to full model tuning.

Tianxiang Sun, Zhengfu He, Hong Qian et al.

Introduces least-to-most prompting, combining problem decomposition and step-by-step solving, greatly enhancing large language models' reasoning on complex tasks.

Denny Zhou, Nathanael Schärli, Le Hou et al.

BlockDFL leverages blockchain to enable fully decentralized P2P federated learning, resisting poisoning, enhancing privacy, and reducing communication overhead.

Zhen Qin, Xueqiang Yan, Mengchu Zhou et al.

Proposes Multi2WOZ dataset and conversational pretraining framework TOD-XLMR, significantly improving cross-lingual transfer for task-oriented dialogue.

Chia-Chien Hung, Anne Lauscher, Ivan Vulić et al.