PyData Dublin Virtual Meetup
Details
Excited to announce a virtual PyData Dublin Meetup for Thursday, 25th March! We'll be using the following link to connect:
AGENDA:
[18:45 - 19:00] Registration, networking, and setup.
[19:00 - 19:35] 'Introduction to the World of Neural Networks' by Ashima Chawla
Modern communications networks are so hard to monitor, analyze and control that monitoring and fault management tasks must be automated. A key factor of this is the ability to classify and identify anomalous operation modes and faults, and so distinguish these cases from normal operation conditions. Anomaly detection of sequence data is difficult and can benefit greatly from the use of AI, machine learning and Artificial Neural Networks (ANNs). Recurrent Neural Networks (RNNs) have proved to be a powerful way of analyzing sequence data and have had great success in language modelling. Normally an artificial neural network is a "black box" that cannot provide answers to the question of why an event sequence was classified as anomalous. A relatively new terminology known as ‘interpretability' of neural network decisions can be of great benefit to network operators. My research focuses on sequential data and with the identification of anomalous sequences (anomaly detection). Our techniques have been applied to two separate application domains, cyber security and telecommunications network management.
Data Scientist Ericsson
Linkedin: https://www.linkedin.com/in/ashimachawla16/
[19:35 - 20:10] 'Machine Learning approaches in personalized medicine and Biomarker discovery' by Raheleh Sheibani Tezerji
One of the challenges in medical sciences is how to translate scientific findings into better clinical results. Information and data are the main input for such a translational process. In the era of postgenomic, public databases such as the Cancer Genome Atlas (TCGA), have contributed to the development of various Omics data sets from 33 cancer types. Omics analysis of these data have facilitated the discovery of effective cancer biomarkers. Aside from biomarker discovery, access to such a big database provides significant insight into understanding biological changes, modifications and functions of established biomarkers in certain patients. We will apply an ensemble classifier (EC) machine learning method, for the discovery of cancer driver candidates (mutations) in primary and metastatic cancer samples from TCGA and dbGAP. The accuracy and performance of the method will be estimated by statistical validation techniques. Top driver genes and novel targets will be predicted, which could be used as potential candidate biomarkers, based on biological classification.
Senior Bioinformatics Scientist at Ludwig Boltzmann Institute Applied Diagnostics
Linkedin: https://www.linkedin.com/in/raheleh-sheibani-tezerji-62239693/
[20:10 - 21:30] Networking, BYOB, and wrap-up.
PyData Dublin Virtual Meetup