arXiv 2016

27,850 papers
In 2016, arXiv hosted 1,469 papers with a strong focus on scalable optimization and Bayesian inference. Key research directions included stochastic gradient MCMC for Bayesian inference and robust decentralized optimization within dynamic networks. There was a significant emphasis on high-dimensional data analysis, particularly regarding sampling in non-convex landscapes and learning from low-sample-size biological datasets. Statistical learning theory also saw advancements in understanding dependent dynamical systems and refining generalization bounds. Technical developments during this year centered on refining optimization algorithms, such as randomized block coordinate methods and the Alternating Direction Method of Multipliers (ADMM) integrated with Nesterov’s extrapolation. Research also explored the intersection of neural networks and partial differential equations, fairness in reinforcement learning, and improving the accuracy of Variational Autoencoders. Prolific contributors included Danielle S. Bassett, Shi Gu, and J. Nathan Kutz, reflecting a broader trend toward integrating complex dynamical systems with statistical learning and high-dimensional sampling techniques.

Top Research Directions

Research directions most represented by arXiv 2016 papers.

Preserving Tiny Features Against Spatial Resolution Degradation 92
Overcoming the Cubic Complexity of Kernel Methods 72
Overcoming the Convexity Constraint in Modern Optimization 71
The Empirical Validation Gap 56
Basins of Attraction in Nonconvex Optimization 55
Mitigating Gradient Variance in Stochastic Optimization 54
Overlapping Community Detection in Complex Networks 52
End-to-End Representation Learning in Natural Language Processing 51
Dynamic Preference Modeling in Interactive Systems 51
Generalizing from Synthetic Data to the Real World 51
Stationary Point Convergence in Nonconvex Optimization 50
Generalizing Provable Control to Nonlinear Systems 49
Bidirectional Sequence Modeling with Recurrent Neural Networks 49
Bridging the Lexical Gap in Information Retrieval 48
Kernel Hyperparameter Sensitivity in Distributional Metrics 47

Top Authors

AuthorPapers
Yoshua Bengio39
Walid Saad38
Bernt Schiele32
Chunhua Shen32
Luc Van Gool32
Mohamed-Slim Alouini32
Shuicheng YAN30
Sergey Levine26
Claire Tomlin25
Aaron Courville24
Alfred O. Hero23
Anton van den Hengel22
Christian Theobalt21
Shie Mannor21
Kyunghyun Cho20
Mario Fritz20
Michael Jordan20
Anurag Kumar19
Bernhard Schölkopf19
Pieter Abbeel19

Impactful Papers

Papers from arXiv 2016 that shaped research clusters.

Foundational

Influential

Papers

Title
Accurate Deep Representation Quantization with Gradient Snapping Layer for Similarity Search
Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model
Accurate principal component analysis via a few iterations of alternating least squares
Accurate principal component analysis via a few iterations of alternating least squares.
Accurate Reconstruction of Finite Rate of Innovation Signals on the Sphere
Accurate, robust and reliable calculations of Poisson-Boltzmann binding energies
Accurate Text Localization in Natural Image with Cascaded Convolutional Text Network
ACD: Action Concept Discovery from Image-Sentence Corpora
ACDC: $α$-Carving Decision Chain for Risk Stratification
A Certifiably Correct Algorithm for Synchronization over the Special Euclidean Group
A Characterization of Constant-Sample Testable Properties
A Characterization of the Non-Uniqueness of Nonnegative Matrix Factorizations
A Character-level Convolutional Neural Network for Distinguishing Similar Languages and Dialects
A Character-level Decoder without Explicit Segmentation for Neural Machine Translation
A Cheap Linear Attention Mechanism with Fast Lookups and Fixed-Size Representations
Achievable Moderate Deviations Asymptotics for Streaming Compression of Correlated Sources

« prev Page 7/557 next »