What we’re about
PMLG formed in late 2016 to help build the eco system for Artificial Intelligence and Machine Learning in Australia
This group is for anyone interested in learning to code practical Machine Learning applications with a strong emphasis on deep learning. We loosely follow the Jeremy Howard Fast.Ai course with a remit of peer learning through sharing with your peers.
Our group covers a wide field of AI and data science techniques which we encourage people to explore by participating in solving real world problems.
All skill levels are welcome but you must be prepared to try your hand at coding and importantly share your learning experience.
Website: https://www.pmlg.org
github: https://github.com/pmlg/pmlg.github.io
Twitter: @PerthMLGroup
Values
In attending PMLG you agree to abide by PMLGs values as defined here https://www.pmlg.org/about
Privacy Statement
Contact information: PMLG does not share contacts in our control with third parties. Please contact us immediately you believe a member is abusing the group contacts or messaging systems.
During meetings and events, we may take photographs or create video recordings at any time in connection with the meetup activities. By attending our meetup, you agree that we can use the photographs or recordings for any purpose and at any time, including on our website, in our publicity or in our publications. If you do not agree please let us know before or during the meeting.
General Advice Warning
The information provided on and through all PMLG media, correspondence and events is general in nature only and does not constitute personal advice. The information has been prepared without taking into account your personal objectives, situation or needs. Before acting on any information on this website you should consider the appropriateness of the information having regard to your objectives, situation and needs. Therefore, before you decide to act on any information, buy any product or keep or cancel a similar product that you already hold. Before making any decision, it is important for you to consider these matters and to seek appropriate technical, legal, tax, and other professional advice.
DISCLAIMER
All statements made by PMLG are made in good faith and we believe they are accurate and reliable. PMLG does not give any warranty as to the accuracy, reliability or completeness of information that is contained in its media, except in so far as any liability under statute cannot be excluded. PMLG, its directors, employees and their representatives do not accept any liability for any error or omission in any PMLG media, correspondence or events or for any resulting loss or damage suffered by the recipient or any other person. Unless otherwise specified, copyright of information provided on PMLG media are the property of PMLG.
Upcoming events (2)
See all- Semantic segmentation for application in very high resolution aerial imagesDUG Technology, West Perth, WA
*** Note this talk will be in DUG's offices in West Perth ***
As a bonus DUG will provide guided tours of Bruce: their immersion cooled supercomputer facility. You will also get a tour of the new DUG Nomad - a mobile, modular, data-centre solution. Tours start at 5.30pm followed by presentation at 6pm. #DUGNOMADSpeaker bio:
Dr. Foivos Diakogiannis, Senior Research Scientist at Data61/CSIRO
presents a new modelling approach in semantic segmentation with application in very high resolution in aerial images and field boundaries.Abstract:
In this talk, Dr Diakogiannis will present his latest work on semantic segmentation, introducing the SSG2 modelling framework (Semantic Segmentation Generation 2) as well as applications of it in field boundary detection using time series of input imagery. He will give emphasis on the attention mechanism and how this works in vision problems.The core of SSG2 lies in its dual-encoder, single-decoder network, which is further augmented by a sequence model. This unique architecture accepts a target image along with a collection of support images. Unlike traditional approaches that rely on single, static images, SSG2 innovates by introducing a "temporal" dimension through a sequence of observables for each static input image. Within these sequence elements, the base model predicts the set intersection, union, and difference of labels from the dual-input images. This allows for more nuanced and accurate segmentation results.
The sequence model then takes the reins, synthesizing the final segmentation mask by aggregating these partial views and filtering out noise at each sequence step. The methodology is inspired by techniques in fields like astronomy and MRI, where multiple observations are utilized to enhance data quality.
It capitalizes on the strong correlation of true signals and the uncorrelated nature of noise across different sequence elements, offering a potential for statistical filtering and thereby reducing error rates.