Machine Learning is fantastically useful: come and learn all you can about it from some local experts!

This will be a great fun event. Please come along (I can’t because I’ll be on leave).

Note the catered mixer afterwards... :)

Cheers

Mike

-- 

Michael Charleston

Theoretical Phylogenetics Group

Organiser of Phylomania conference 2018 November 21-23

Associate Professor in Bioinformatics

Co-director of Data, Knowledge and Decisions University Research Theme

Associate Head of School (Learning and Teaching)

School of Natural Sciences

University of Tasmania

AUSTRALIA

Phone: +61 3 6226 2444

From: Anya Reading <anya.reading@utas.edu.au>
Date: Thursday, 27 September 2018 at 10:28 am
To: Michael Charleston <michael.charleston@utas.edu.au>
Subject: DKD EVENT on THUR OCT 4 at IMAS: DataTas Seminar Series: Practical Machine Learning

View this email in your browser

DATA TAS SEMINAR SERIES
2:00pm - 5:00pm October 4, 2018 // IMAS Waterfront Building (Aurora)

Hello DataTasmanians,

We're pleased to announce our next Seminar Series on Machine Learning to be held 2:00pm - 5:00pm October 4 in the Aurora Lecture Theatre, IMAS waterfront building. DataTas is partnering with the UTAS Research Theme for Data, Knowledge, Decisions (DKD) to bring you this exciting joint event.

This session, we have Professor Anya Reading, the leader of the Compute Earth group (UTAS) giving us an introduction to Machine Learning and its application for research. We also have Matilda Brown (UTAS) to talk to us about Support Vector Machines (SVMs) in ecological research and Ben Schroeter (IMAS, BoM) to give us some practical tips on tuning Artificial Neural Networks (ANNs).

After the session we will hold an informal gathering in the Flex Space (catered!) where you can chat with the speakers and other ML enthusiasts, and where we will discuss the formation of a Machine Learning community of practice in Hobart.

See you there!

DataTas

Learning from data:
A partnership of mind and machine
Anya Reading (UTAS)

The presentation provides an overview and introduction to the use of Machine Learning in research.  Anya will outline the main approaches to using 'the machine’ to learn from data, unsupervised and supervised learning, including the advantages and limitations of ML algorithms in current use.  She will show how to gain research insight from ML outputs using examples from the Compute Earth research group, and show how the strengths of both humans and computers may be used together for best results.

Anya Reading came to UTAS in 2007.  She leads the Compute Earth group which pioneers geophysical data collection in remote or challenging locations such as Antarctica and outback Australia.  This adventuring spirit extends to exploring and extending the ways that we can use computers to learn from data, and also to initiatives such as Art-Science data visualisation.  She is Professor of Geophysics, in Physics at UTAS Sandy Bay, with strong connections to Earth Sciences and IMAS.

Support Vector Machines in Ecology
Matilda Brown (UTAS)

Matilda will describe some of the practical uses of machine learning – in particular, Support Vector Machines (SVMs) in ecological studies. Matilda will give us a brief introduction to SVMs, and go on to describe some of their applications in her PhD. This includes pixel classification in image analysis, environmental and taxonomic identification from the processed images, and detection of ecological changes in the fossil record.

Matilda Brown is a PhD student in Biological Sciences at the University of Tasmania. She is working with Greg Jordan, Tim Brodribb and Barbara Holland to find new ways of applying Machine Learning techniques to answer palaeoecological questions. Matilda has spent most of her adult life trying to get people excited about leaves - either as a bushwalking guide, demonstrator or academic. She is interested in making life easier for other biologists and investigating the ecological and evolutionary history of our flora. 

Tuning the Machine:
Practical tips for training Artificial Neural Networks
Ben Schroeter (IMAS, Bureau of Meteorology)

Artificial Neural Networks (ANNs) provide a convenient way to model highly nonlinear systems. With the emergence of plug-and-play libraries (i.e. TensorFlow) they are more accessible than ever but remain to many a black box with a few dials and buttons for tuning. Through his masters research into the application of Machine Learning to rainfall estimation, Ben will share some practical tips on how to tune an ANN and offer suggestions on how to avoid some of the common pitfalls, such as overfitting and suboptimal minima. No idea what those mean? That's OK! Ben will give you a crash course and some ideas to help you tune your model.

Ben Schroeter is a Support Scientist at the Bureau of Meteorology working on the next generation of city-scale Numerical Weather Prediction (NWP) models, and PhD candidate at the Institute for Marine and Antarctic Studies working on Antarctic NWP. Ben enjoys tackling challenging computer science problems and melting CPUs/GPUs with parallelised code.

@datatas

datatasau

datatas.com

datatas

Email

JOIN DATATAS

Become a full member of DataTas for the cost of a couple of cups of coffee and get access to exclusive member benefits.

Student: $5/yr     non-Student: $10/yr
(Registration available at all our events)

Copyright © 2018 DATA TAS, All rights reserved.
You are receiving this email because you opted in at our website or provided your details to us at one of our events.

Our mailing address is:

DATA TAS

Tasmanian University Union, PO Box 5055, University of Tasmania LPO

Sandy Bay, Tas 7005

Australia

Add us to your address book

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.

University of Tasmania Electronic Communications Policy (December, 2014).
This email is confidential, and is for the intended recipient only. Access, disclosure, copying, distribution, or reliance on any of it by anyone outside the intended recipient organisation is prohibited and may be a criminal offence. Please delete if obtained in error and email confirmation to the sender. The views expressed in this email are not necessarily the views of the University of Tasmania, unless clearly intended otherwise.