Previous Next
Scroll down to
explore more

LIONlab seminars

By members and visitors

Learning from examples in optical imaging

May 12, 2016
Speaker: Prof. Demetri Psaltis, EPFL
Abstract: Optical tomography has been widely investigated for biomedical imaging applications. In recent years optical tomography has been combined with digital holography and has been employed to produce high-quality images of phase objects such as cells. We describe a method for imaging 3D phase objects in a tomographic configuration implemented by training an artificial neural network to reproduce the complex amplitude of the experimentally measured scattered light. The network is designed such that the voxel values of the refractive index of the 3D object are the variables that are adapted during the training process.


Independent Component Analysis

April 27, 2016
Speaker: Marco Zugliani, uniTN intern at LIONlab
Abstract: Independent component analysis (ICA) is a method for separating an observed set of signal mixtures into a set of statistically independent component signals, or source signals. The goal of ICA is to find a linear representation of non-Gaussian data so that the components are statistically independent, or as independent as possible. This technique has different application, e.g. image noise reduction, face recognition, medical signal processing.


Graph Partitioning Problem

April 20, 2016
Speaker: Tahir E. Kalayci, postdoctoral research fellow at LION lab - University of Trento, Italy
Abstract: Graphs are frequently used by computer scientists as abstractions when modelling an application problem and cutting a graph into smaller pieces is one of the fundamental algorithmic operations. Even if the final application concerns a different problem (such as traversal, finding paths, trees, and flows), partitioning large graphs is often an important sub-problem for complexity reduction or parallelism. There are different approaches like global optimization, iterative improvement heuristics, multilevel graph partitioning, evolutionary methods and further meta-heuristics for solving graph partitioning problem. In this talk, we are going to focus on bi-partitioning of the graphs and introduce heuristic based algorithms for solving this problem.


scikit-learn, machine learning for the 99.9%

April 13, 2016
Speaker: Fabian Pedregosa, postdoctoral fellow at Chaire Havas-Dauphine / INRIA Paris
Abstract: scikit-learn is a machine learning library for Python. Although nowadays largely a community effort, the library was born from the effort of a small number of developers. In this talk I will describe the library vision, goals, and identify key aspects that made it possible to transition from a small number of developers to the current ~100 active contributors. The talk will finish with a brief hands-on introductory tutorial.


Parallel stochastic optimization for machine learning

April 6, 2016
Speaker: Fabian Pedregosa, postdoctoral fellow at Chaire Havas-Dauphine / INRIA Paris
Abstract: With its cheap per-iteration cost, stochastic gradient descent has become in recent years the workhorse of huge-scale optimization and is behind the success of recent AI achievements such as deep learning. At the same time, practical parallel variants of stochastic gradient descent have been developed. These algorithms achieve a theoretical linear speedup in the number of cores under strong assumptions, such as sparsity of gradients. After a brief review of the current literature, I will present our main contribution, the ASAGA algorithm, a fully parallel version of the incremental gradient algorithm SAGA that (unlike plain stochastic gradient descent) enjoys fast linear convergence rates. Furthermore, we prove that ASAGA can obtain a theoretical linear speed-up on multicore systems under weaker assumptions that previous literature. We present results of an implementation on a 40-core architecture illustrating the practical speed-up as well as the hardware overhead. This is joint work with my colleagues at INRIA-Paris RĂ©mi Leblond and Simon Lacoste-Julien.


Approximated Nearest Neighbors

March 16, 2016
Speaker: Andrea Mariello, PhD student at LIONlab - University of Trento, Italy
Abstract: The problem of finding the nearest neighbor or the first K neighbors of a point occurs very often in various machine learning and pattern recognition applications. For instance, in a classification task we can load all points in a dataset, with their class labels, at first, and then we can assign a new point (the query point) to the class of its nearest neighbor or the class voted by the majority of its first K neighbors. Since this model is lazy, lacking a preprocessing phase, we have to start a new search for each query point. When we increase the number of points in the dataset as well as the number of queries and data dimensionality, finding exact neighbors of all query points by a linear search on the dataset translates to an overall complexity of O(d N^2), where d is the number of dimensions and N the number of points. If our problem is not bound to exact solutions, we can significantly improve the search speed by using Approximated Nearest Neighbors (ANN) techniques. In this seminar, we are going to introduce one of these techniques, which is based on hashing and k-means clustering, that achieves a complexity of O(d N^(3/2)).


Exact solutions of pickup and delivery problems: comparative analysis of different models

March 7, 2016
Speaker: Elvina Gindullina, visiting PhD student from Ufa State Aviation Technical University, Russia
Abstract: The goal of routing problem ONE-TO-ONE type (or Traveling Salesman Problem with Pickup and Delivery (TSPPD)) is the delivery of goods from producers to consumers by the shortest route in such manner that goods from every producer must be delivered to specific consumers. We construct a few linear integer or mixed integer formalizations of such problems where the number of constraints grows polynomially with the number of points. The standard optimization programs can be used directly for such formalizations. Computational experiments in the program CPLEX 12.6 were organized for comparing the proposed formalizations.


LIONbusiness workshop: nail the "LION way" strategy

An outlook on disruptive technologies

The LION business workshop covers the LION vision , with training by "data rock stars" and pioneers, and with clear business examples including big data, data mining and predictive analytics, actionable models and optimization. This powerful combination of tools is at the basis of prescriptive analytics.

The workshop is delivered in two flavors:

  • LIONbusiness I: one-day full immersion intended for managers and technical people. No math or software experience required, we focus on vision, business examples, disruptive innovation made possible by connecting data to models and decisions ( prescriptive analytics )
  • LIONbusiness II: two-days version with hands-on experience. The first day is the same as version I. The second day is dedicated to hands-on practice with software tools, therefore suggested for people with technical competence and passion for using software. The objective is to make you fully capable of using LION tools to solve your business problems.

Both versions are open to a maximum of seven participants, to allow for ample discussion time and focus on your selected business interests.
The standard location of LIONbusiness is Venice, alternative locations are Verona and Trento (the barycenter of Europe in the Italian Alps).
Depending on availability, the LIONbusiness workshop can be delivered on-premises at your company site.

LION is a complex array of mechanisms, like the engine in an automobile, but you do not need to know the inner workings of the engine in order to realize its tremendous benefits. Knowing how to drive a company in the LION way requires obtaining a clear vision of the possibilities, the goals and the ways to reach them. We promise you will get this vision and understand how to realize its impact in your company in our single or two-days workshop.

The LION way (Learning and Intelligent Optimization) combines learning from data and optimization to improve dynamic business processes. It deals with increasing the automation level and connecting data directly to improved decisions and actions. Therefore, it goes beyond traditional business intelligence, towards predictions and directly towards more automation in improving decisions.

Our promise: four gold nuggets of empowerment for improving services, products and ROI:

Machine learning plus otimization.
 

Realizing the vision of extreme automation.

Self-service insight from big data.
 

Unlock the hidden value in your business data.

Predictive and actionable analytics.
 

What-if scenarios, improve decisions and ROI.

Groundbreaking innovation in services and products
 

Realize the dream of continuous creation by self-reflective tools.

Past LION events

A sample of LION-related events and locations in the last years.

  • LION 8 Conference

    Gainesville, Florida - USA, Feb 16-21, 2014
  • LION 6 Conference

    Paris, Jan 16-20, 2012
  • LION 5 Conference

    Rome, Jan 17-21, 2011
  • LION 4 Conference

    Venice, Jan 18-22, 2010

LIONbusiness topics and schedule

LIONbusiness I: one-day full immersion

An intensive and concrete introduction, intended for managers and technical people alike. Vision, goals, strategies, discussion. No math. No software.
Standard schedule is from 9.00 to 18.00 with two coffee breaks, lunch and workshop dinner.
Given the strictly limited participation, in addition to the presentations by top experts and pioneers, important values of our workshops are ample time for questions, in-depth discussions and brainstorming about specific business problems .

Topics of the first day are:

Worst fears and best potential of LION (learning and intelligent optimization).

Let's discuss about the overall LION vision and power but also about potential pitfalls, resistances, proper ways to get organized to implement the LION way in your business.

Big data, predictive analytics and optimization (prescriptive analytics)

Huge amounts of data are produced during business operations. The LION way develops methods and tools to mine this treasure and extract actionable insight. Applications are far-reaching, ranging from marketing and e-commerce, to finance, bioinformatics, healthcare, engineering, and social networks. After models through "learning from data" are available, automated improvement motors run continuously to obtain better and better solutions.

Deep machine learning

The increasing availability of huge amounts of data in machine readable format represents an unprecedented opportunity but also a formidable challenge. The most recent advances in deep machine learning are needed to scale to large datasets, and deal with unstructured data.

Models which can be explained to people

Models able to provide interpretable explanations for decision makers are especially appealing. Providing human explanation facilitates implementation, debugging, and adoption.

How to realize the potential of powerful multiple-objective optimization without deep math knowledge

Remember, you do not need to be a mechanical engineer to drive your car. But you need knowledge of the destination, the dashboard, steering wheel, etc.

Creating a self-service analytics and optimization ecosystem

The refreshing feeling of being liberated by intermediate layers of "data priests" to get exactly what you want from a well-orchestrated system cannot be explained in words. You must live through it.

LIONbusiness II: two-days version with hands-on experience

The second day, in addition to LIONbusiness I, is designed for people with some technical competence in and passion for using software tools. The day is dedicated to hands-on practice with tools on concrete and paradigmatic business problems.

Topics of the second day are:

Filtering data and identifying relevant information and strategic goals

Machine learning and big data are powerful, but it is up to you to define goals, select relevant data, filter out noise, embed the strategy into your daily operations.

Selecting the best model and tool for selected applications

Classification, clustering, risk assessment, quality control, marketing and sales. Feel free to specify your core interest so that we will include it into the test-drive experience.

Hands-on experience on your problems

You are very welcome to come with a sample of your data to test our tools on.

LIONbusiness rates (tuition)

 
Fees are regular rates, please inquire with us for group discounts, on premises sessions and next available dates. Workshop language: English or Italian.

LIONbusiness I: one-day full immersion
 
EURO 900 (per participant)
LIONbusiness II: two-days, with hands-on practice
 
EURO 1500 (per participant)
LIONbusiness I: on premises
 
EURO 1200 (per participant, plus travel and lodging)
LIONbusiness II: on premises
 
EURO 2000 (per participant, plus travel and lodging)

Contact us to pre-register

Get ready for the "LION way" and select the best option for you and your business.

Send us a message to fix details

Contact Info

Address: DISI - Università of Trento,
Via Sommarive 5, 38123, Trento

Email: mauro.brunato((AT)) unitn.it

Our place (Trentino Dolomites)

img