Konstantina Dritsa retweeted
Tool Augmented Language Models.
The #AI4Code reading group is also back in July with Aaron Parisi, #Google researcher and lead author of the #TALM paper👇
arxiv.org/pdf/2205.12255.pdf
Konstantina Dritsa retweeted
Grateful to @GreencityzenIoT for hosting an incredible plenary meeting in Marseille, France last week. 👏🏼 Thanks to all project partners for their active participation and insightful discussions. Exciting times ahead, stay tuned for updates! 🌍🚀
Konstantina Dritsa retweeted
Plan is to throw a party in the Andromeda galaxy 1B years from now. Everyone welcome, except for those who litter
Tune in for today's presentation!
ΑΙ at AUEB talk, Tue 22 Nov 16:15-17:00, "What we Talk about When we Talk about Politics", Dritsa, Pavlopoulos, Louridas (accepted at NeurIPS 2022). Room T106, Troias building and virtually via MS Teams.
teams.microsoft.com/l/meetup…
Really excited to share that our #dataset of the Greek Parliament proceedings for computational linguistics and political analysis was accepted to #NeurIPS2022! Congratulations to the whole team! @KaitiThoma @itpavlopoulos @louridas Paper and code coming soon!
Konstantina Dritsa retweeted
The stars are aligning...so to speak. ⭐
Our @NASAWebb team has fully aligned the telescope's primary imager with its mirrors, keeping its optics on track to meet or exceed science goals. Tune in at noon ET (16:00 UTC) for an update: go.nasa.gov/3IhPOvK
#UnfoldTheUniverse

ALT While the purpose of this image was to focus on the bright star at the center for alignment evaluation, Webb's optics and NIRCam are so sensitive that the galaxies and stars seen in the background show up. At this stage of Webb’s mirror alignment, known as “fine phasing,” each of the primary mirror segments have been adjusted to produce one unified image of the same star using only the NIRCam instrument. This image of the star, which is called 2MASS J17554042+6551277, uses a red filter to optimize visual contrast. Credit: NASA/STScI
Konstantina Dritsa retweeted
GPT-3 can now make changes to existing content, not just predict what comes next. Released in the API today: openai.com/blog/gpt-3-edit-i…
Together with G. Nikitopoulos, @louridas and @d1m1tr0 we created CrossVul, a cross-language vulnerability dataset enriched with commit data. I am happy to share that our work will appear at @FSEconf. Preprint: dimitro.gr/assets/papers/NDL…
Konstantina Dritsa retweeted
Οφειλόμενη απαντηση σ'όσους διαδίδουν ότι η συγκεντρωση στο εφετείο ευθύνεται για την έξαρση του κορωνοιού: Σύμφωνα με επίσημα στοιχεία η έξαρση του ιού τις δυο τελευταίες εβδομάδες είναι πτωτική στην Αττική και αυξητική στην υπόλοιπη επικράτεια. ΣΤΑΜΑΤΗΣΤΕ ΝΑ ΛΕΤΕ ΨΕΜΑΤΑ.
Konstantina Dritsa retweeted
further on software reliability, we have been working with T. Sotiropoulos, S. Chaliasos and @CoolSWEng on "a model for detecting faults in build specifications". Now I am very happy to share that our work will appear at OOPSLA / @splashcon! Preprint: dimitro.gr/assets/papers/SCM…
Konstantina Dritsa retweeted
#happeningnow: Konstantina Dritsa from AUEB presenting the neural networks zoo, at the PRACE Training Course on Machine Learning in HPC.
events.prace-ri.eu/event/994…
#GRNET #HPC #MachineLearning #PRACE
Konstantina Dritsa retweeted
An Experimental Document & Snippet Retrieval Search Engine for the CORD-19 (COVID-19) dataset, based on our best BioASQ7 system. #COVID19 @allen_ai @BioASQ
Online demo: cslab241.cs.aueb.gr:5000/
Papers: nlp.cs.aueb.gr/pubs/aueb_at_… nlp.cs.aueb.gr/pubs/emnlp201…
Code: github.com/nlpaueb/aueb-bioa…
Konstantina Dritsa retweeted
A really smart approach for identifying malicious access to a user's account by converting their mouse movements to an image and then feeding it to a CNN.
splunk.com/en_us/blog/securi…
Konstantina Dritsa retweeted
🔴We, at @iMEdDLab, created an app to track the geographical distribution of #COVID19 in Greece and around the world. Try the beta, in Greek >> lab.imedd.org/covid19/
#covid19Gr #COVID2019 #κορονοιος #menoumespiti
@iMEdD @andefined @Troboukis @gsarig
🔴COVID-19: 31 νέα κρούσματα | 418 συνολικά | 79 νοσηλεύονται - 13 διασωληνωμένοι
Παρακολουθήστε την εξάπλωση της νόσου στην Ελλάδα και στον υπόλοιπο κόσμο μέσα από τον διαδραστικό χάρτη που ανέπτυξε το @iMEdD Lab.
lab.imedd.org/covid19/
Konstantina Dritsa retweeted
New from Google Research! REALM: realm.page.link/paper
We pretrain an LM that sparsely attends over all of Wikipedia as extra context. We backprop through a latent retrieval step on 13M docs. Yields new SOTA results for open domain QA, breaking 40 on NaturalQuestions-Open!
Lived to see the picture of #takistsoukalas in a paper about neural networks (fig.1) arxiv.org/pdf/2002.00883.pdf #ANTEGEIA #mlgreece
Konstantina Dritsa retweeted
Some folks still seem confused about what deep learning is. Here is a definition:
DL is constructing networks of parameterized functional modules & training them from examples using gradient-based optimization.... facebook.com/722677142/posts…
Not all heroes wear capes!
Review #1: Weak Reject
Review #2: Weak Reject
Review #3: Reject
No one:
AC: Ok. I read the reviews and the rebuttal. All the concerns are addressed. ACCEPT.
This is an AC that only exists in your wildest dreams.
#ICLR2020
openreview.net/forum?id=rJg7…
3
Konstantina Dritsa retweeted
wonderful news: our work with T. Sotiropoulos and @CoolSWEng on practical fault detection in Puppet programs will appear at the @ICSEconf (main track)! Preprint already available: dimitro.gr/assets/papers/SMS… #Puppet #DevOps
Konstantina Dritsa retweeted
With PDGuard your clients and your organization can ensure that personal data stay private and is used only as authorized. This happens by design and in an auditable manner. medium.com/@dimitris.i.mitro…


























