Dropout significantly reduces catastrophic forgetting in neural networks, with experimental data showing 15% better retention of old tasks and 20% improved adaptation to new tasks.

Ian J. Goodfellow, Mehdi Mirza, Da Xiao et al.

Proposes spectral and local clustering graph networks with input-size-independent parameters, effective on low-dimensional graphs.

Joan Bruna, Wojciech Zaremba, Arthur Szlam et al.

Introduces reparameterization and auto-encoding variational Bayes (AEVB) for efficient inference in continuous latent variable models.

Diederik P Kingma, Max Welling

Gradient-based visualization of CNN class models and saliency maps enhances interpretability and weakly supervised object localization.

Karen Simonyan, Andrea Vedaldi, Andrew Zisserman

Proposes FFT-based convolution acceleration, achieving over 10x speedup in training and inference for deep CNNs.

Michael Mathieu, Mikael Henaff, Yann LeCun

Deep Mixture of Experts excels on MNIST, learning location and class-specific experts.

David Eigen, Marc'Aurelio Ranzato, Ilya Sutskever

Proposes a two-point function value-based zero-order convex optimization algorithm achieving near-gradient convergence rates.

John C. Duchi, Michael I. Jordan, Martin J. Wainwright et al.

Proposed a quantum algorithm for simulating sparse Hamiltonians with exponential precision improvement.

Dominic W. Berry, Andrew M. Childs, Richard Cleve et al.

Introducing geometric Hamiltonian Monte Carlo (HMC) with SoftAbs metric for efficient sampling in hierarchical models, overcoming funnel distribution issues.

M. J. Betancourt, Mark Girolami

Dynamic Mode Decomposition (DMD) analyzes nonlinear systems via linear operator eigendecomposition, enhancing computational efficiency and noise mitigation.

Jonathan H. Tu, Clarence W. Rowley, Dirk M. Luchtenburg et al.

Using Quincke rotation to induce self-propulsion, the study demonstrates hydrodynamic interactions stabilize a polar-liquid phase in dilute colloidal populations, suppressing giant density fluctuations.

Antoine Bricard, Jean-Baptiste Caussin, Nicolas Desreumaux et al.

This paper derives non-asymptotic oracle inequalities for LASSO in high-dimensional VAR models, ensuring prediction and estimation accuracy with high probability.

Anders Bredahl Kock, Laurent A. F. Callot

Proposes an optimal composition theorem for differential privacy, using hypothesis testing and information inequalities to define tight privacy bounds.

Peter Kairouz, Sewoong Oh, Pramod Viswanath

Analyzed convergence of decentralized gradient descent with fixed step size, achieving O(1/k) rate and linear convergence under strong convexity.

Kun Yuan, Qing Ling, Wotao Yin

Distributed reactive power control via ADMM reduces power losses and maintains voltage stability in photovoltaic-integrated grids.

Petr Šulc, Scott Backhaus, Michael Chertkov

Proposes a game-theoretic double auction mechanism improving storage unit utility by 130.2%.

Yunpeng Wang, Walid Saad, Zhu Han et al.

Proposed an efficient training method for NADE models, enhancing density estimation performance.

Benigno Uria, Iain Murray, Hugo Larochelle

DeCAF features extracted from ImageNet-trained CNN outperform traditional features on multiple vision tasks, achieving over 20% accuracy improvements.

Jeff Donahue, Yangqing Jia, Oriol Vinyals et al.

Turney’s dual-space model unifies analogy and composition through domain/function cosine similarities; the supplied text omits numerical scores.

Peter D. Turney

Survey on data-driven grasp synthesis methods, categorized by known, familiar, and unknown objects.

Jeannette Bohg, Antonio Morales, Tamim Asfour et al.