UNISURF unifies neural implicit surfaces and radiance fields for mask-free multi-view 3D reconstruction, achieving high accuracy.

Michael Oechsle, Songyou Peng, Andreas Geiger

Proposes EW and MW metrics to separately evaluate language models' syntactic systematicity and behavior, revealing overestimation of their capabilities.

Benjamin Newman, Kai-Siang Ang, Julia Gong et al.

Proposes a universal algorithm for constructing equivariant MLPs for arbitrary matrix groups, enabling efficient layer design for complex symmetries.

Marc Finzi, Max Welling, Andrew Gordon Wilson

Proposes TransFuser, a Transformer-based multi-modal fusion model, reducing collisions by 76% in urban driving scenarios.

Aditya Prakash, Kashyap Chitta, Andreas Geiger

LATTICE leverages multimodal content to mine latent item relationships via graph learning, boosting recommendation accuracy by 12% on key metrics.

Jinghao Zhang, Yanqiao Zhu, Qiang Liu et al.

Introduces an actuator-aware kino-dynamic planner and hierarchical control enabling humanoid robot flips and spins in simulation.

Matthew Chignoli, Donghyun Kim, Elijah Stanger-Jones et al.

Detect out-of-distribution data for pretrained Transformers using contrastive loss and Mahalanobis distance, achieving near-perfect performance.

Wenxuan Zhou, Fangyu Liu, Muhao Chen

Constrained language models achieve few-shot semantic parsing, significantly outperforming baselines.

Richard Shin, Christopher H. Lin, Sam Thomson et al.

Introduces CLIPScore, a reference-free metric correlating 74% with human judgments, surpassing CIDEr and SPICE.

Jack Hessel, Ari Holtzman, Maxwell Forbes et al.

Introduces the concept of knowledge neurons using integrated gradients, identifying neurons responsible for factual knowledge in BERT, enabling knowledge editing without fine-tuning, with an average of 4.13 neurons per fact.

Damai Dai, Li Dong, Yaru Hao et al.

Introduced Domain Conditional Pointwise Mutual Information to address surface form competition, enhancing zero-shot task accuracy.

Ari Holtzman, Peter West, Vered Shwartz et al.

Proposes Condenser, a pre-training architecture with conditional dense representations, significantly improving retrieval tasks.

Luyu Gao, Jamie Callan

ZeRO-Infinity leverages heterogeneous memory to support models with hundreds of trillions of parameters without code refactoring.

Samyam Rajbhandari, Olatunji Ruwase, Jeff Rasley et al.

SR3 achieves image super-resolution via iterative denoising, reaching a 50% fool rate on CelebA-HQ 8x tasks.

Chitwan Saharia, Jonathan Ho, William Chan et al.

Retrieval-augmented architecture reduces hallucination by over 60%, outperforming SOTA on two knowledge-grounded tasks, enhancing factual accuracy in dialogue.

Kurt Shuster, Spencer Poff, Moya Chen et al.

Integrating source and target context modeling, including coreference resolution and multi-hop reasoning via Transformer-XH, improves detection of previously fact-checked claims by over 10 MAP points.

Shaden Shaar, Firoj Alam, Giovanni Da San Martino et al.

SummScreen dataset for abstractive screenplay summarization, challenging due to dialogue and plot details.

Mingda Chen, Zewei Chu, Sam Wiseman et al.

GridToPix trains policies in gridworlds with terminal rewards, achieving SPL 0→64 and success 1%→25% in Embodied AI tasks.

Unnat Jain, Iou-Jen Liu, Svetlana Lazebnik et al.

AR-LSAT combines Transformer and symbolic reasoning, revealing deep reasoning gaps in current models with accuracy below 35%. ARM achieves 65%, still behind humans.

Wanjun Zhong, Siyuan Wang, Duyu Tang et al.

Addresses few-shot image generation overfitting with cross-domain distance consistency loss and anchor strategy.

Utkarsh Ojha, Yijun Li, Jingwan Lu et al.