Home
 
Home

Understanding how intelligence works is largely recognized as one of the most formidable endeavors of science to date with virtually unlimited technological fallout.

By acknowledging the key role played by learning in the development of intelligent systems, we investigate the foundations of learning and learning algorithms from the viewpoints of computer science, mathematics, and statistics.

Starting from our long standing experience in the design of algorithms for extracting visual information from images, our efforts are directed toward computer vision applications.

Leveraging on close collaborations with molecular biologists, clinicians, and radiologists we broadened our interests in the areas of medical image analysis and computational biology.

 

 
libMsFeatS

(Model assessment for Feature Selection) is a platform independent C++ toolbox that includes the implementation of elastic net and group LASSO for binary and multi-class classification problems. For the latter, the (groups of) features that are selected are the ones that best discriminate between all the classes simultaneously.
Our toolbox also provides model selection functionalies, allowing the user to automatically select the best performing regularization parameters. The input data file is compliant with libSVM (http://www.csie.ntu.edu.tw/~cjlin/libsvm/) format, to favor the interchange with classification.

The libMsFeatS homepage with code and instructions can be found here

 
PPlus

PPlus is a simple environment to execute Python code in parallel on many machines without much effort. It is actually a fork of Parallel Python, another simple but powerful framework for parallel execution of python code, which lacks features needed for effective use in our daily research.

The PPlus homepage can be found here together with the source code.

 
L1L2Signature

L1L2Signature is an implementation of an unbiased framework originally thought for gene expression analysis. The framework was used in many real applications and the package is composed by a set of Python scripts and a set of useful classes and functions that could be used to manually read and/or analyze high-throughput data extending/integrating the proposed pipeline.
The L1L2Signature homepage can be found here together with the source code.

 
L1L2Py

implements the l1l2 regularization with double optimization described and studied in [1,2]. The package is implemented in Python(≥2.5.0) and requires NumPy(≥1.3.0).

The L1L2Py homepage can be found here together with the source code.