
Vasily Konovalov retweeted
🚨 New paper alert! 🚨
Our new paper on ArXiv: "ImageReFL: Balancing Quality and Diversity in Human-Aligned Diffusion Models".
It tackles a key challenge in diffusion models: aligning with human preferences without collapsing diversity
👉 arxiv.org/abs/2505.22569
1/5
Vasily Konovalov retweeted
#GSoC + @anasedova = new relation extraction model in #DeepPavlov Library! @michael_galkin check this out -
link.medium.com/iJU7w3QAzjb
Vasily Konovalov retweeted
#2021 is awesome! For the second time in a year, there is reason to be proud 🎉
DeepPavlov reached 250k downloads 🔥 Thank you our Enthusiastic & Diverse Community 🌏
💙 #DeepPavlov! #OpenSource #Conversational #AI Framework 🎅
Vasily Konovalov retweeted
#2021 is awesome! We are so proud! 🤩
⭐️ We got 5k stars on #GitHub
💙 #DeepPavlov! #OpenSource #Conversational #AI Framework 🎅
Vasily Konovalov retweeted
A long-overdue blog post about commonsense reasoning for NLP (based on our ACL 2020 tutorial): veredshwartz.blogspot.com/20…
Vasily Konovalov retweeted
Hurray, guys! ✌
Our #DREAM team passed the certification of #AlexaPrize Socialbot Grand Challenge 4 💪🎉
Vasily Konovalov retweeted
DeepPavlov Library 0.14.0 release link.medium.com/jI7CT7Oxvcb
Vasily Konovalov retweeted
I finally watched all the talks I wanted to, ended up importing 56 papers to my bib, and now present to you:
🎉 My 13 favorite papers (sorted alphabetically) at #EMNLP2020! 🔥
[1/15]
Vasily Konovalov retweeted
I'll give a talk on multiskill conversational AI @ NVIDIA GTC conference and @danielko will showcase our #Deeppavlov tools for AI assistants followed by live session. Happy to answer your questions :) nvidia.com/en-us/gtc/session…
Vasily Konovalov retweeted
Starting from the latest release #DeepPavlov 🎅 supports both top deep learning frameworks – TensorFlow and PyTorch.
⚡ PyTorch is available for several DP models including BERT-based classifier, NER, ranking and text QA.
#nlproc
https://nitter.gq/t.co/oggu0W1EwJ
Vasily Konovalov retweeted
DeepPavlov🎅0.12.0 is out📣
🔥PyTorch Support-long awaited!
👨🎤👨🔧👨🚀Multi-task BERT
🛠Basic RASA Configs Support for Go-Bot
⛓Entity Linking
🏆top_n Answers API Update in ODQA
🕵️♀️Hybrid NER Models Trained on OntoNotes
💾BoolQ Dataset Reader
Notes In Our Blog🚀deeppavlov.ai/blog/tpost/58y…
Vasily Konovalov retweeted
The Bias-Variance Trade-Off & "DOUBLE DESCENT" 🧵
Remember the bias-variance trade-off? It says that models perform well for an "intermediate level of flexibility". You've seen the picture of the U-shape test error curve.
We try to hit the "sweet spot" of flexibility.
1/🧵
Vasily Konovalov retweeted
An excellent introduction to the Transformer architecture, the workhorse of all BERT, GPT-X, etc. Very happy to see this replacing the phrase-based LMs I worked on back in 2005-07, ushering in a new era for machine translation. youtube.com/watch?v=iDulhoQ2…
Vasily Konovalov retweeted
New! Do you use NaturalQuestions, TriviaQA, or WebQuestions? It turns out 60% of test set answers are also in the train set. More surprising, 30% of test questions have a close paraphrase in the train set. What does it mean for models? Read arxiv.org/abs/2008.02637 to find out! 1/
Vasily Konovalov retweeted
The 10th Lisbon Machine Learning School (LxMLS) has just started! Link: lxmls.it.pt Live streaming here: tiny.cc/LxMLS2020. #LxMLS2020 @istecnico
Vasily Konovalov retweeted
The recordings from #EEML2020 are now online at eeml.eu/program: lectures, tutorials, panel, firesides. The coding resources from the tutorials are on github github.com/eemlcommunity/Pra….
Enjoy the summer with AI!
Vasily Konovalov retweeted
A kind user on Github has put together this tremendous list of courses for AI and machine learning. Topics range from introductory lectures to advanced lectures.
#MachineLearning #AI
buff.ly/2WBAup8
Vasily Konovalov retweeted
How GPT3 works. A visual thread.
A trained language model generates text.
We can optionally pass it some text as input, which influences its output.
The output is generated from what the model "learned" during its training period where it scanned vast amounts of text.
1/n
Vasily Konovalov retweeted
Still cropping and modifying BERT diagrams from Devlin et al. (2019)? I spent several hours redrawing BERT in PowerPoint so you don't have to... Perfect for use in presentations, papers, etc.! Happy to share under Releasing under CC BY 4.0 cs.uwaterloo.ca/~jimmylin/BE…
Vasily Konovalov retweeted
New blog post up: so, you've probably seen all the tweets about GPT-3.
GPT-3 is objectively a step forward in the field of AI text-generation, but the current hype on VC Twitter misrepresents the model's current capabilities. GPT-3 isn't magic. minimaxir.com/2020/07/gpt3-e…
























