Interested in a PhD in the area of #ResponsibleAI?
We have two fully-funded PhD scholarships at @NortheasternLDN and @Unikent.
Join us to work on the verification (https://t.co/jvObJGvWsT) and testing (https://t.co/ec9aUI1KWF) of deep learning models.
Apply by 13 March 2023!
Excited for this opportunity to discuss my research at the intersection of computer systems and machine learning and the lessons learnt while working with fantastic researchers at @LSDSImperial, @graphcoreai, and @NCHLondon.
#systems, #DeepLearning
For our first Expeditions in Experiential AI seminar of 2022, we are joined by @akoliousis to discuss combating performance and accuracy trade-offs in deep learning systems. Register today to join us on February 16 at 1pm ET.
#letsAItogether
https://t.co/M9VTcOeKn5
Graphcore Research has developed a new #BERT model that uses grouped transformations.
#GroupBERT enables IPU users to reduce training time by 50%, while retaining the same level of accuracy.
https://t.co/1go1z9Nn8y
@TDataScience#NLP#AI
Artificial Intelligence and Philosophy: Why Do They Need Each Other?
Join us on 9th April 5pm (BST) for this free stimulating and thought-provoking roundtable discussion by leading experts in the field of #AI, #Philosophy, and #Ethics.
Click to register: https://t.co/0rkjPdXLDm
Join Senior Lecturer in Computer Science, Dr Alexandros Koliousis, for our free Computer Science Taster Lecture on The Natural Language Processing Grand Prix on 3 March 2021 at 13:00 (GMT)
Register here: https://t.co/knZ1T5yQRY
Excited that our blog post on Training philosopher engineers, co-authored with my colleague Brian Ball from @NCHLondon, found a home at The Pebble at @PeaSoupEthics. https://t.co/PLUX9c2voy
@akoliousis presenting our @LSDSImperial work on #Crossbow at #VLDB2019. Crossbow is a multi-GPU deep learning system that uses small batch sizes yet scales well. Code on GitHub https://t.co/ggBreKhtNR and TF integration forthcoming
Deep learning needs more theory.
Many workshops I have helped organize in the last decade (e.g. at IPAM) have attempted to bring ML/DL folks together with mathematicians and... https://t.co/jvOH36lYH8
We have released #Crossbow, a multi-#GPU#deeplearning system that scales despite using small batch sizes: https://t.co/ggBreKhtNR #Crossbow uses synchronous model averaging. Check out the paper (with quotes by @JeffDean and @ylecun): https://t.co/kvCyrAB3gF Feedback welcome!