Giorgos Paraskevopoulos

PhD student. Speech and Language Processing group at NTUA @NtuaSlp

[ILSP - Athena RC @ILSP_AthenaRC]

📢Now available the much anticipated and greatly needed: 🆕Meltemi (/meltémi/), the first open-source Greek LLM trained by #ILSP of @athenaRICinfo . Read more: https://t.co/kr4qFKXTUh

[Giorgos Paraskevopoulos @georgepar_91]

@angeloskath Great demo!
The original tweet for more info is here: https://t.co/3s6Sh2Nn4L

[ILSP - Athena RC @ILSP_AthenaRC]

📢Now available the much anticipated and greatly needed: 🆕Meltemi (/meltémi/), the first open-source Greek LLM trained by #ILSP of @athenaRICinfo . Read more: https://t.co/kr4qFKXTUh

[Efthymis Georgiou @efthygeo]

Happy to share our work "M3: Multimodal Masking applied to Sentiment Analysis" which we presented at Interspeech 2021, Brno.

Paper: https://t.co/p6htP5Xa7c

Code: https://t.co/fxiITWln1n

[Constantinos Karouzos @ckarouzos]

📢 Super excited to share that our paper "UDALM: Unsupervised Domain Adaptation through Language Modeling" got into @NAACLHLT #NAACL2021 🥳🥳 Joint work @NtuaSlp @ecentua w/ @georgepar_91 and Prof. @apotam ! #NLProc

📄 Paper: https://t.co/7mMtQ4iaKF

Summary thread follows ⬇️🔜

[Giorgos Paraskevopoulos @georgepar_91]

Happy to share our paper "Multiresolution and Multimodal Speech Recognition with Transformers", which was accepted in ACL 2020. https://t.co/c5uormGpaK

We used dot product attention for visual conditioning in multimodal ASR

This is the product of my internship in Amazon Lab126.

[TiaGia @tyiannak]

My latest medium article on emotion and behavior-aware speech analytics: "Can AI improve the way you speak?" https://t.co/LWQ6ZzOduG #voicefirst #conversationalAI

[Thomas Wolf @Thom_Wolf]

🔥Pytorch-Transformers 1.0🔥

Six NLU/NLG architectures: BERT, GPT, GPT-2, Transfo-XL, XLNet, XLM
Total: 27 pretrained models

Still the same
-Superfast onboarding
-SOTA scripts: GLUE, SQuAD, Text generation

New
-Unified API
-Access hidden-states, attentions...
-Torchscript
-...

[Giorgos Paraskevopoulos @georgepar_91]

Congrats @giannis_daras on the awesome presentation. There is some exciting work ahead :)

[Giannis Daras @giannis_daras]

Slides from my talk at #spaCyIRL regarding sparse attention factorizations are available here: https://t.co/KFZR2KRFdz …
Thanks for the massive interest, paper and code are going to be released soon.
Slides partially describe joint work with: @georgepar_91 @AlexGDimakis @apotam

[Giannis Daras @giannis_daras]

Appreciate that you liked it! 💥😁 I will also share the slides of the presentation soon!

[AI at Meta @AIatMeta]

We’ve developed a tool that applies natural language processing (#NLP) and information retrieval techniques directly to source code text, in order to produce a #machinelearning-based code search system.

https://t.co/4FcMU5Xwza

[Kawin Ethayarajh @ethayarajh]

When and why does king - man + woman = queen? In my #ACL2019 paper with @DavidDuvenaud and Graeme Hirst, we explain what conditions need to be satisfied by a training corpus for word analogies to hold in a GloVe or skipgram embedding space. 1/4

blog: https://t.co/2TUnbBwnx5

[Filippos Kokkinos @filippos_kok]

Interested in burst photography or learnable optimization schemes??? Drop by at our poster #169 at #CVPR2019 to see how we developed an iterative NN for burst restoration based on large-scale optimization techniques.

https://t.co/CG2RJFps1x

[Giannis Daras @giannis_daras]

I was training pytorch model on multiple gpus, getting out of memory due to single gpu loss computation. This amazing gist written by @Thom_Wolf is a nice and clean workaround. Check it out, if you haven't already.
https://t.co/qbhL4k1BmZ

[ntua-slp @NtuaSlp]

We are happy to announce that our short paper "Attention-based Conditioning Methods for External Knowledge Integration" will appear at @ACL2019_Italy ! Congrats @katemargatina @cbaziotis @apotam 🎉🎉🎉 #acl2019nlp #NLProc

[Alexandra Chronopoulou @alexandraxron]

Happy for @katemargatina!!!! 🎉🎉🎉Her very interesting work on different methods of including linguistic information in attention mechanisms has been accepted at @ACL2019_Italy! 🥳 Well done!!!

[Christos Baziotis @cbaziotis]

The pre-print of our #naacl2019 paper is now online.
"SEQ^3: Differentiable Sequence-to-Sequence-to-Sequence Autoencoder for Unsupervised Abstractive Sentence Compression" (https://t.co/9rXfS7H8HV). With @ionandrou, Ioannis Konstas and @apotam.
Code: https://t.co/YyaphxtI8f

[Derek Chen @derekchen14]

Just watch it already.

[Jan Leike @janleike]

How do we uncover failures in ML models that occur too rarely during testing? How do we prove their absence?

Very excited about the work by @DeepMindAI’s Robust & Verified AI team that sheds light on these questions! Check out their blog post:

https://t.co/P7PqQULDPs

[AUEB NLP Group @AUEBNLPGroup]

Next meeting, Tue 2 April, 17:15-19:00: @alexandraxron presenting "An Embarrassingly Simple Approach for Transfer Learning from Pretrained Language Models" (NAACL 2019) + Discussion of Peters et al. "To Tune or Not to Tune? Adapting..." (https://t.co/yZWkMe491O). Room A36.

[Rachel Thomas @math_rachel]

Repository to track progress in NLP, including datasets and current state-of-the-art for the most common NLP tasks. @seb_ruder https://t.co/ObNIzYYss4