I am excited to announce our @iclr_conf 2023 paper: "A VAE for Transformers with Nonparametric Variational Information Bottleneck" https://t.co/2LIMCaxMJF We propose to model Transformer embeddings as nonparametric mixture distributions using Dirichlet processes. @FabioFehr

Multiple PhD positions in deep learning for NLP and neuroscience are available in my Natural Language Understanding group at Idiap, Switzerland @idiap_ch ( https://t.co/Jr76w2gkkZ). Good salaries, beautiful location, and a world-class AI research community. #NCCR_EvolvingLanguage

I'm really excited about our new work on modelling the latent space of Transformers as nonparametric mixture distributions, and defining a Nonparametric Variational AutoEncoder with a Transformer encoder-decoder. Paper to be presented at #ICLR: https://t.co/5VI5SzW7Ce

Idiap Research Institute has several attractive open positions to head research groups in AI, with a preference for cross-disciplinary and industry-driven research. Consideration of applications will start soon. : https://t.co/KWwT9rM9Wr

Our new paper (w/ @JamieBHenderson): A novel way of encoding syntactic structure for SRL task by inputting graph relations directly into the attention mechanism of Transformer: Syntax-Aware Graph-to-Graph Transformer for Semantic Role Labelling, https://t.co/Y2Elz6X9LD @Idiap_ch

Come and work with us! We have seven permanent jobs and one postdoc! Check us out!

Our Findings of @emnlp2020 paper “Graph-to-Graph Transformer for Transition-based Dependency Parsing” (w/ @JamieBHenderson) is now available: @Idiap_ch @EPFL Arxiv: https://t.co/97kRn1mi5G Presentation: https://t.co/OD01C4cr9Y Code:

2. "Multilevel text alignment with cross-document attention" (w/ @nlpxuhui, @nlpnoah). 3. "Plug and play autoencoders for controlled text generation" (w/, @fmai, @ivanspmontero, @nlpnoah, @JamieBHenderson). More details coming soon. (2/2)

Our paper (w/ @JamieBHenderson) titled 'Graph-to-Graph Transformer for Transition-based Dependency Parsing' has been accepted as a long paper in Findings of @emnlp2020 . #emnlp2020 updated preprint and code will be available soon.

The announcent and call for papers for EACL 2021 is now online. Plan to submit, plan to attend!! See you there.

Senior postdoctoral position open in my group (3+years). Some French required. You'll be involved in a visible computational linguistics group and in the new exciting Evolving language project! Check us out. For all information see announcement at https://t.co/AKheyhlMc5

Grateful that my #acl2020nlp theme paper "The Unstoppable Rise of Computational Linguistics in Deep Learning" was accepted. Traces how the nature of language has and will impact neural network architectures, including variable binding in Transformer.

I am excited to share that our paper, "End-to-End Bias Mitigation by Modelling Biases in Corpora", is accepted to #acl2020 #acl2020nlp #nlproc Joint work with @boknilev and @JamieBHenderson

A Postdoc and multiple PhD positions are available in my Natural Language Understanding group at Idiap, Switzerland @idiap_ch ( https://t.co/P7BcO5b8hb). Good salaries, beautiful location, and a world-class AI research community. #NCCR_EvolvingLanguage

We have great new SoTA syntactic parsing results with RNG Transformer: Recursive Non-Autoregressive Graph-to-Graph Transformer for Dependency Parsing with Iterative Refinement, https://t.co/qTBovee4Jv. Retrains BERT to correct syntactic structures, graph2graph.

We stand by Italy during these trying times. Share your Support for our Italian friends, They are our colleagues, friends and family. Cari amici, siamo con voi. #COVID19 #WeStandWithItaly

Great seeing the launch of Scotland's leading AI startup - Alana @alanathebot spinning out from @HeriotWattUni and the development of their Conversational AI Interfaces Congrats to @oliverlemon and @verena_rieser and team on the launch

The border between Geneva and France. I feel safer already.