• Publications
  • Influence
Assessing the accuracy of prediction algorithms for classification: an overview
We provide a unified overview of methods that currently are widely used to assess the accuracy of prediction algorithms, from raw percentages, quadratic error measures and other distances, and
Bayesian surprise attracts human attention
  • L. Itti, P. Baldi
  • Computer Science, Psychology
    Vision Research
  • 5 December 2005
A Bayesian framework for the analysis of microarray expression data: regularized t -test and statistical inferences of gene changes
TLDR
A Bayesian probabilistic framework for microarray data analysis is developed that derives point estimates for both parameters and hyperparameters, and regularized expressions for the variance of each gene by combining the empirical variance with a local background variance associated with neighboring genes.
Searching for exotic particles in high-energy physics with deep learning.
TLDR
It is shown that deep-learning methods need no manually constructed inputs and yet improve the classification metric by as much as 8% over the best current approaches, demonstrating that deep learning approaches can improve the power of collider searches for exotic particles.
Prediction of protein stability changes for single‐site mutations using support vector machines
TLDR
The method can accurately predict protein stability changes using primary sequence information only, it is applicable to many situations where the tertiary structure is unknown, overcoming a major limitation of previous methods which require tertiary information.
SCRATCH: a protein structure and structural feature prediction server
SCRATCH is a server for predicting protein tertiary structure and structural features. The SCRATCH software suite includes predictors for secondary structure, relative solvent accessibility,
Improving the prediction of protein secondary structure in three and eight classes using recurrent neural networks and profiles
TLDR
Ensembles of bidirectional recurrent neural network architectures, PSI‐BLAST‐derived profiles, and a large nonredundant training set are used to derive two new predictors for secondary structure predictions, and confusion matrices are reported.
Autoencoders, Unsupervised Learning, and Deep Architectures
  • P. Baldi
  • Mathematics, Computer Science
    ICML Unsupervised and Transfer Learning
  • 2 July 2011
TLDR
The framework sheds light on the different kinds of autoencoders, their learning complexity, their horizontal and vertical composability in deep architectures, their critical points, and their fundamental connections to clustering, Hebbian learning, and information theory.
A principled approach to detecting surprising events in video
  • L. Itti, P. Baldi
  • Computer Science
    IEEE Computer Society Conference on Computer…
  • 20 June 2005
TLDR
A new theory of sensory surprise is presented, which provides a principled and computable shortcut to important information and develops a model that computes instantaneous low-level surprise at every location in video streams.
...
1
2
3
4
5
...