Search Results for author:
Found 32 papers, 9 papers with code
In Your Pace: Learning the Right Example at the Right Time
Initially, we define the difficulty of a training image using transfer learning from some competitive "teacher" network trained on the Imagenet database, showing improvement in learning speed and final performance for both small and competitive networks, using the CIFAR-10 and the CIFAR-100 datasets.
Relearning Forgotten Knowledge: on Forgetting, Overfit and Training-Free Ensembles of DNNs
An extensive empirical evaluation with modern deep models shows our method's utility on multiple datasets, neural networks architectures and training schemes, both when training from scratch and when using pre-trained networks in transfer learning.
United We Stand: Using Epoch-wise Agreement of Ensembles to Combat Overfit
Our method allows for the incorporation of useful knowledge obtained by the models during the overfitting phase without deterioration of the general performance, which is usually missed when early stopping is used.
Pruning the Unlabeled Data to Improve Semi-Supervised Learning
In the domain of semi-supervised learning (SSL), the conventional approach involves training a learner with a limited amount of labeled data alongside a substantial volume of unlabeled data, both drawn from the same underlying distribution.
Semi-Supervised Learning in the Few-Shot Zero-Shot Scenario
Semi-Supervised Learning (SSL) is a framework that utilizes both labeled and unlabeled data to enhance model performance.
The Dynamic of Consensus in Deep Networks and the Identification of Noisy Labels
This introduces a problem when training in the presence of noisy labels, as the noisy examples cannot be distinguished from clean examples by the end of training.
Active Learning Through a Covering Lens
Deep active learning aims to reduce the annotation cost for the training of deep models, which is notoriously data-hungry.
Active Learning on a Budget: Opposite Strategies Suit High and Low Budgets
Investigating active learning, we focus on the relation between the number of labeled examples (budget size), and suitable querying strategies.
Ranked #1 on
Active Learning
on CIFAR10 (10,000)
The Grammar-Learning Trajectories of Neural Language Models
These findings suggest that there is some mutual inductive bias that underlies these models' learning of linguistic phenomena.
Principal Components Bias in Over-parameterized Linear Models, and its Manifestation in Deep Neural Networks
Empirically, we show how the PC-bias streamlines the order of learning of both linear and non-linear networks, more prominently at earlier stages of learning.
More Is More -- Narrowing the Generalization Gap by Adding Classification Heads
Our model can be employed during training time only and then pruned for prediction, resulting in an equivalent architecture to the base model.
Boosting the Performance of Semi-Supervised Learning with Unsupervised Clustering
We also notably improve the results in the extreme cases of 1, 2 and 3 labels per class, and show that features learned by our model are more meaningful for separating the data.
Multiclass non-Adversarial Image Synthesis, with Application to Classification from Very Small Sample
In the small-data regime, where only a small sample of labeled images is available for training with no access to additional unlabeled data, our results surpass state-of-the-art GAN models trained on the same amount of data.
Generative Latent Implicit Conditional Optimization when Learning from Small Sample
GLICO learns a mapping from the training examples to a latent space and a generator that generates images from vectors in the latent space.
Multi-Modal Deep Clustering: Unsupervised Partitioning of Images
Simultaneously, the same deep network is trained to solve an additional self-supervised task of predicting image rotations.
Ranked #6 on
Image Clustering
on Tiny-ImageNet
Let's Agree to Agree: Neural Networks Share Classification Order on Real Datasets
We further show that this pattern of results reflects the interplay between the way neural networks learn benchmark datasets.
On The Power of Curriculum Learning in Training Deep Networks
We address challenge (i) using two methods: transfer learning from some competitive ``teacher" network, and bootstrapping.
Theory of Curriculum Learning, with Convex Loss Functions
We also prove that when the ideal difficulty score is fixed, the convergence rate is monotonically increasing with respect to the loss of the current hypothesis at each point.
Coming to Your Senses: on Controls and Evaluation Sets in Polysemy Research
This and additional results point to the conclusion that performance gains as reported in previous work may be an artifact of random sense assignment, which is equivalent to sub-sampling and multiple estimation of word vector representations.
Gaussian Mixture Generative Adversarial Networks for Diverse Datasets, and the Unsupervised Clustering of Images
In order to provide a better fit to the target data distribution when the dataset includes many different classes, we propose a variant of the basic GAN model, called Gaussian Mixture GAN (GM-GAN), where the probability distribution over the latent space is a mixture of Gaussians.
Curriculum Learning by Transfer Learning: Theory and Experiments with Deep Networks
We provide theoretical investigation of curriculum learning in the context of stochastic gradient descent when optimizing the convex linear regression loss.
Distance-based Confidence Score for Neural Network Classifiers
The reliable measurement of confidence in classifiers' predictions is very important for many applications and is, therefore, an important part of classifier design.
Outta Control: Laws of Semantic Change and Inherent Biases in Word Representation Models
This article evaluates three proposed laws of semantic change.
Hidden Layers in Perceptual Learning
Studies in visual perceptual learning investigate the way human performance improves with practice, in the context of relatively simple (and therefore more manageable) visual tasks.
Every Untrue Label is Untrue in its Own Way: Controlling Error Type with the Log Bilinear Loss
Deep learning has become the method of choice in many application domains of machine learning in recent years, especially for multi-class classification tasks.
Novelty Detection in MultiClass Scenarios with Incomplete Set of Class Labels
Our method is based on the initial assignment of confidence values, which measure the affinity between a new test point and each known class.
Optimized Linear Imputation
Since most data analysis and statistical methods do not handle gracefully missing values, the first step in the analysis requires the imputation of missing values.
Topic Modeling of Behavioral Modes Using Sensor Data
A common use of accelerometer data is for supervised learning of behavioral modes.
A Cheap System for Vehicle Speed Detection
The reliable detection of speed of moving vehicles is considered key to traffic law enforcement in most countries, and is seen by many as an important tool to reduce the number of traffic accidents and fatalities.
Online Learning in The Manifold of Low-Rank Matrices
When learning models that are represented in matrix forms, enforcing a low-rank constraint can dramatically improve the memory and run time complexity, while providing a natural regularization of the model.
Beyond Novelty Detection: Incongruent Events, when General and Specific Classifiers Disagree
We define a formal framework for the representation and processing of incongruent events: starting from the notion of label hierarchy, we show how partial order on labels can be deduced from such hierarchies.
