Applying normalized DTW (nDTW) for path similarity evaluation improves human correlation and RL training in navigation tasks.

Gabriel Ilharco, Vihan Jain, Alexander Ku et al.

LakhNES enhances multi-instrument music generation using cross-domain pre-training with Transformer and NES-MDB dataset.

Chris Donahue, Huanru Henry Mao, Yiting Ethan Li et al.

Introduces Sparse Momentum for training sparse neural networks from scratch, achieving up to 5.61x speedup while maintaining dense performance levels.

Tim Dettmers, Luke Zettlemoyer

BAM employs multi-task distillation with teacher annealing, achieving >1.5% improvement on GLUE benchmarks over standard methods.

Kevin Clark, Minh-Thang Luong, Urvashi Khandelwal et al.

Introduces Product Keys-based structured memory layer, enabling billion-parameter capacity with efficient exact search, outperforming deeper transformers in language modeling.

Guillaume Lample, Alexandre Sablayrolles, Marc'Aurelio Ranzato et al.

Proposes a collaborative referring task under continuous, partially observable context, with a large-scale dataset of 6,760 dialogues to evaluate neural models.

Takuma Udagawa, Akiko Aizawa

Proposes a particle filter-based active perception framework for robot model uncertainty estimation, improving manipulation accuracy.

Oliver Kroemer, Scott Niekum, George Konidaris

Using three GP algorithms and random search, constructing two compact features achieves comparable or superior performance to full feature sets across 21 datasets, with GP-GOMEA showing best results.

Marco Virgolin, Tanja Alderliesten, Peter A. N. Bosman

Prototypes-guided, model-agnostic counterfactuals significantly improve speed and interpretability, reducing search time by 77% and increasing local interpretability metrics IM1 by 13%.

Arnaud Van Looveren, Janis Klaise

Combines heuristic A* search and B-spline optimization for fast, safe quadrotor trajectory generation in complex environments.

Boyu Zhou, Fei Gao, Luqi Wang et al.

Proposes a multi-dataset mixing approach with scale and shift-invariant loss for robust monocular depth estimation, achieving superior zero-shot cross-dataset transfer on five diverse datasets.

René Ranftl, Katrin Lasinger, David Hafner et al.

Proposes a GNN-based spectral clustering method avoiding costly spectral decomposition, enhancing graph pooling performance.

Filippo Maria Bianchi, Daniele Grattarola, Cesare Alippi

Introduced a deep active learning method with adaptive acquisition, showing superior performance across datasets.

Manuel Haussmann, Fred A. Hamprecht, Melih Kandemir

Hierarchical Optimal Transport (HOTT) combines topic models and word embeddings to efficiently measure document similarity, outperforming WMD in speed with comparable accuracy.

Mikhail Yurochkin, Sebastian Claici, Edward Chien et al.

Proposes Multilevel Graph Convolutional Neural Network (MGCN) for molecular property prediction, achieving MAE of 0.0129 eV on QM9.

Chengqiang Lu, Qi Liu, Chao Wang et al.

Saliency-based interpretation method reveals word alignments in NMT, outperforming attention weights with up to 20% AER reduction, applicable offline and online.

Shuoyang Ding, Hainan Xu, Philipp Koehn

Unified framework using Markov blanket and free energy to describe multiscale physical systems' self-organization.

Karl Friston

Using bit extraction, ReLU networks achieve r/d<p≤2r/d; fixed width is nearly optimal, while periodic activations yield near-exponential rates.

Dmitry Yarotsky, Anton Zhevnerchuk

Introducing VICES as an action space for deep RL, enhancing sample efficiency, safety, and transferability in contact-rich tasks.

Roberto Martín-Martín, Michelle A. Lee, Rachel Gardner et al.

LV-Net combines two-stream pyramid and nested encoder-decoder for robust salient object detection in optical RSIs, outperforming 14 SOTA methods on the ORSSD dataset.

Chongyi Li, Runmin Cong, Junhui Hou et al.