Interested in doing a PhD IT? Large number of scholarships and projects available this year at Deakin. deakin.edu.au/information-te…

springer.com/journal/521/up…

My recent talk at the Human aligned reinforcement learning workshop at ICDL2021 focusing on Broad-XAI. m.youtube.com/watch?v=WYPBk6…

Our latest paper presenting a framework for developing broad based explainable reinforcement learning has been published. Also provides an extensive review of current progress towards developing such XAI. link.springer.com/article/10.100…

Great new paper from our team with an approach that apologises and adjusts its behaviour based user reactions. An example of using impact minimisation with implicit user-based Pareto policy selection in multi-objective RL (MORL) Hadassah Harland @amp1874 link.springer.com/article/10.100…

Open letter calls for Ed Husic, Anthony Albanese and Department of Industry, Science and Resources to support AI safety 🇦🇺 Signed by Toby Ord, Simon Goldstein, M A Osborne, Richard Dazeley, Daniel Murfet, @amp1874, Hadassah Harland, Hunter Jay and many others Ange Lavoipierre Mike Cherney australiansforaisafety.com.au

Nice video on human-in-the-loop AI and it’s potential shortcomings. aus01.safelinks.protection.outlook.com/?url=https%3A%…

Ep7 of AGI Show podcast is out! We interview Prof of AI & ML Deakin University, Richard Dazeley (Richard Dazeley) on: * What AGI may look like * How we can respond on tech & governance * His research into multi-objective RL & its applications to AGI alignment theagishow.com/2082819/133415…

Great new paper from our research group. Congratulations Haddie on this expansive review of AI Apology in Artificial Intelligence Review. Peter Vamplew, [email protected], Frankie Cruz link.springer.com/article/10.100…

OMG I agree with tony jones. #QandA

#qanda the bigots seriously want us to feel sorry for them? They aren't the ones being discriminated against.

If every decision made is evidence based, why did your government ditch the carbon tax? #QandA

Have u read the Koran Pauline? I doubt you can read or could understand the language it is written in. QandA

LNP rely on market forces because that allows them to do nothing. QandA

Paid hazzlewood and they pocketed it and shut down. QandA

No trains to deer park from ballarat between 3:52 and 7:11. Then usually wait at Ballan for 30 min. Useless service. #59minuteBallarat

Train to bus service doesn't connect. Have to wait 30 min for bus to uni. Then bus goes in circles doubling trip time. #59minuteBallarat

Must catch 6:27am train from deer park to get to feduni by 9 (153 min). Driving only takes 65. May as well have no trains. #59minuteBallarat