• Publications
  • Influence
Large Scale Online Learning of Image Similarity through Ranking
TLDR
OASIS is an online dual approach using the passive-aggressive family of learning algorithms with a large margin criterion and an efficient hinge loss cost, which suggests that query-independent similarity could be accurately learned even for large-scale datasets that could not be handled before. Expand
Large Scale Online Learning of Image Similarity Through Ranking
TLDR
OASIS is an online dual approach using the passive-aggressive family of learning algorithms with a large margin criterion and an efficient hinge loss cost, which suggests that query independent similarity could be accurately learned even for large scale data sets that could not be handled before. Expand
Information Bottleneck for Gaussian Variables
TLDR
A formal definition of the general continuous IB problem is given and an analytic solution for the optimal representation for the important case of multivariate Gaussian variables is obtained, in terms of the eigenvalue spectrum. Expand
Euclidean Embedding of Co-occurrence Data
TLDR
This paper describes a method for embedding objects of different types, such as images and text, into a single common Euclidean space, based on their co-occurrence statistics, and shows that it consistently and significantly outperforms standard methods of statistical correspondence modeling. Expand
Learning from Noisy Large-Scale Datasets with Minimal Supervision
TLDR
An approach to effectively use millions of images with noisy annotations in conjunction with a small subset of cleanly-annotated images to learn powerful image representations and is particularly effective for a large number of classes with wide range of noise in annotations. Expand
Max-margin Classification of Data with Absent Features
TLDR
It is shown how incomplete data can be classified directly without any completion of the missing features using a max-margin learning framework, and it is shown that the elegant handling of missing values by this approach allows it to both outperform other methods when the missing values have non-trivial structure, and be competitive with other method when the values are missing at random. Expand
Reduction of Information Redundancy in the Ascending Auditory Pathway
TLDR
Information about stimulus identity was somewhat reduced in single A1 and MGB neurons relative to single IC neurons, when information is measured using spike counts, latency, or temporal spiking patterns, but this difference was due to differences in firing rates. Expand
Encoding Stimulus Information by Spike Numbers and Mean Response Time in Primary Auditory Cortex
TLDR
In auditory cortex, whereas spike counts carry only partial information about stimulus identity or location, the additional availability of relatively coarse temporal information is sufficient in order to extract essentially all the sensory information available in the spike discharge pattern, at least for the relatively short stimuli commonly used in auditory research. Expand
A unifying principle underlying the extracellular field potential spectral responses in the human cortex.
TLDR
A new global, cross-frequency (10-100 Hz) neuronal process is reflected in a significant reduction of the power spectrum slope of the ECoG signal. Expand
An Online Algorithm for Large Scale Image Similarity Learning
TLDR
The non-metric similarities learned by OASIS can be transformed into metric similarities, achieving higher precisions than similarities that are learned as metrics in the first place, suggesting an approach for learning a metric from data that is larger by orders of magnitude than was handled before. Expand
...
1
2
3
4
5
...