Deep Learning
On the Effective Number of Linear Regions in Shallow Univariate ReLU Networks: Convergence Guarantees and Implicit Bias
ReLU neural networks with a single hidden layer in a binary classification setting.
OpenOOD: Benchmarking Generalized Out-of-Distribution Detection Jingkang Y ang
From the problem setting perspective, OOD detection is closely related to neighboring fields including anomaly detection (AD), open set recognition (OSR), and model uncertainty, since methods developed for one domain are often applicable to each other.
Improved Fine-Tuning by Better Leveraging Pre-Training Data
As a dominant paradigm, fine-tuning a pre-trained model on the target data is widely used in many deep learning applications, especially for small data sets.
Supplemental Material: Debiased Self-Training for Semi-Supervised Learning
We further explore the training stability of FixMatch when using pre-trained models on various tasks. Figure 2 illustrates several failure cases of FixMatch.
Arch - A Conv3x3 Conv1x1 Conv3x3 Modelling NN as a Information Computation Graph Ops Across Archs Ops Across Positions (Arch - A)
In these cases, the NN architecture itself can be viewed as data and needs to be modeled. A better modeling could help explore novel architectures automatically and open the black box of automated architecture design.
Deep Combinatorial Aggregation
Neural networks are known to produce poor uncertainty estimations, and a variety of approaches have been proposed to remedy this issue.
MobTCast: Leveraging Auxiliary Trajectory Forecasting for Human Mobility Prediction
It handles both the temporal and semantic contexts.