[July 8, 10:00] Cliff Click: Fairly techy H2O innards

Datalab is proud to announce the second lecture of the Summer Camp 2015. After Kate Smith Miles who gave an excellent presentation on benchmarking of optimization algorithms - cornerstone of machine learning, our special guest will be Cliff Click, CTO at 0xdata.

Place: T9:105 (New Building of CTU)

Cliff will talk about the architecture of H2O system for fast scalable machine learning. You can also ask him what does the sparkling water mean.

About the Speaker

Read more: [July 8, 10:00] Cliff Click: Fairly techy H2O innards

[July 3, 10:00] Kate Smith-Miles: Understanding strengths and weaknesses of optimization algorithms with new visualization tools and methodologies

Datalab is proud to announce the first lecture of Summer Camp 2015. Our special guest will be Kate Smith-Miles, Professor of Mathematical Sciences at Monash University, Australia.

Place: T9:111 (New Building of CTU)

Objective assessment of optimization algorithm performance is notoriously difficult, with conclusions often inadvertently biased towards the chosen test instances. Rather than reporting average performance of algorithms across a set of chosen instances, we discuss a new methodology to enable the strengths and weaknesses of different optimization algorithms to be compared across a broader instance space. Results will be presented on TSP, timetabling and graph coloring to demonstrate:

  1. how pockets of the instance space can be found where algorithm performance varies significantly from the average performance of an algorithm;
  2. how the properties of the instances can be used to predict algorithm performance on previously unseen instances with high accuracy;
  3. how the relative strengths and weaknesses of each algorithm can be visualized and measured objectively; and
  4. how new test instances can be generated to fill the instance space and provide desired insights into algorithmic power.

About the Speaker

Read more: [July 3, 10:00] Kate Smith-Miles: Understanding strengths and weaknesses of optimization...

[May 4, 14:30] Tomáš Kuzin: Sensor Failure Detection and Prediction

Place: TH:A-1347

Sensors are collecting huge amounts of data, but besides the state of observed phenomena itself, what else can we determine from the collected measurements? Aim of this presentation is to show some methods currently used to detect anomalies in sensory data sets, which may suggest potential sensor failure and use such a knowledge for predictive maintenance.

Presentation slides are available here and video here.

[April 20, 14:00] Martin Barus: Bayesian Networks

Place: TH:A-1347

Bayesian networks, also known as belief networks, belong to the family of probabilistic graphical models. Combining principles from graph theory, probability theory, computer science, and statistics, BNs are used for reasoning under uncertainty as they infer conditional probability distributions of unobserved variables based on the evidence set (observed variables) and joint probability distribution. They have been successfully applied in victim identification, fraud detection, spam filtering, cancer risk modeling etc.

Presentation slides are available here and video here.

Follow Us

Copyright (c) Data Science Laboratory @ FIT CTU 2014–2016. All rights reserved.