Check out our #AutoML methodology for efficient #superresolution DNNs and the associated TinyPerceptualSR model at #ECCV2020 this week. Efficient model design for perceptually high quality super-resolution is now possible.

Complementary cat photos for more attention to CFP of ACM Distributed ML workshop, co-located with #CoNEXT 2020 🔉🔉🔉 Submission due on September 11th 📜📚✏️Both technical and position papers up to 6-page long are solicited! More info 🔗: distributedml.org

Looking forward to the #mobicom2020 conf., starting today. Check out our work, "SPINN: Synergistic Progressive Inference of Neural Networks over Device and Cloud", this Wednesday. Paper: dl.acm.org/doi/10.1145/33… Pitch: youtu.be/Vzs_VU9K0b0 Presentation: youtu.be/wmCNVt-NQd0

Join us on 01/12 in our DistributedML workshop. Registration open! conferences2.sigcomm.org/co-next/2020/#… Thrilled to announce our program (distributedml.org/program/), featuring great keynotes and an intriguing panel discussion. Looking forward to a great event! #CoNEXT2020 #ACM #MachineLearning

Continuing our work on multi-exit DNNs, we present PersEPhonEE, an *on-device personalisation* system for DNNs at #HotMobile'21. PersEPhonEE alleviates the lack of labelled data through a self-distillation technique and reduces the cost of on-device training by more than 20x.

First paper of the day at ACM #HotMobile2021: "It's always personal: Using Early Exits for Efficient On-Device CNN Personalisation" dl.acm.org/doi/10.1145/34… Ilias Leontiadis Steve Laskaridis Stylianos I. Venieris nic lane @samsung

Exciting news! Our proposal for the 1st Workshop on Security & Privacy for Mobile AI (MAISP) has been accepted ACM MobiSys 2021. With Kleomenis Katevas, Ilias Leontiadis, Cristian Canton. It will be hosted online this year! Call-For-Papers coming soon! Powered by: Telefónica Scientific Research,@facebookai,Samsung Electronics

Consider submitting and attending MobiSys 2021 workshops! Thanks to the online format this year we cover a broad spectrum of emerging and relevant topic!

Do you work on Security/Privacy aspects of ML on mobile or edge devices? Check our Call for Papers for the 1st Workshop on Security & Privacy for Mobile AI (MAISP), collocated with ACM MobiSys 2021 Deadline: May 7th More info: maisp.gitlab.io #dataprivacy #cybersecurity

7 days to submit your paper at the 1st Workshop on Security & Privacy for Mobile AI (MAISP), ACM MobiSys 2021. If you work on Security/Privacy of ML on mobile/edge/network devices, check out the info: maisp.gitlab.io Great keynotes & program! #dataprivacy #cybersecurity

Due to several requests, we are extending the MAISP deadline ACM MobiSys to 14/5. If you work on Security/Privacy of ML on mobile/edge/network devices, check it out: maisp.gitlab.io Ilias Leontiadis,Kleomenis Katevas,Cristian Canton Great keynotes/program! concordia-h2020.eu ACCORDION Project H2020

1st Workshop on Security & Privacy for Mobile AI (MAISP) this Thursday, 14:00 CEST. Collocated in Mars ACM MobiSys 2021! With Ilias Leontiadis, Kleomenis Katevas and Cristian Canton Exciting program & keynotes: @paulpatras, Hamed, Diego Perino, nic lane maisp.gitlab.io

Join us today at the 1st Workshop on Security & Privacy for Mobile AI (MAISP) in Mars (also virtually) at 14:00 CEST. Very exciting program & keynotes: Patras Paul, Hamed, Diego Perino, nic lane With Nicolas Kourtellis, Ilias Leontiadis and Cristian Canton. maisp.gitlab.io/program/

Don't miss Steve Laskaridis's talk today at Federated Learning One World Seminar (FLOW) about our work to support federated learning on a wide variety of heterogeneous devices (from weakest to the most powerful). Full paper available here: arxiv.org/abs/2102.13451

Last edition of the DistributedML.org workshop was a great success with great papers and speakers (online!). This year the deadline is on September 17, co-located with Conext. Consider submitting your work! See you there Steve Laskaridis @tforcworc #ML #DistributedML #FL

Very excited about our paper, titled "Smart at what cost? Characterising Mobile DNNs in the wild" being accepted at IMC'21. It includes insights from large scale in-the-wild benchmarking of DNNs in the most popular apps of Play Store across devices with gaugeNN. Coming out soon..

We are proud to announce our speaker and panelist lineup for DistributedML (co-located with #conext21), which includes: Virginia Smith (CMU), Mosharaf Chowdhury (Michigan), Chuanxiong Guo (ByteDance), Bo Li (UIUC), Vijay Janapa Reddi (Harvard) and Tianyi Chen (RPI). New submission deadline 24/09

Tune in this Thursday (Session 15) for the presentation of our work at #IMC2021, titled "Smart at what cost?: Characterising mobile DNNs in the wild". Mario Almeida will be presenting our insights for the state of DNN deployment in the wild, across top "AI-powered" apps in Play Store.

Two weeks to go for #EMDL'22 deadline, to be co-located with #MobiSys 2022! Submit your best work on embedded and mobile deep learning! Website: emdl22.github.io Submissions: emdl22.hotcrp.com

Samsung Electronics AI in Cambridge has open positions for ML Engineers. Today we released an advertisement for our Distributed AI group, check out a preview before it hits the website: drive.google.com/file/d/1HtOI6I… We also have engineering roles in our ML group and Speech/Audio group. Join us!