Rohit Pillai – Medium

Rohit Pillai
in
Towards Data Science

CODIR: Train smaller, faster NLP models

A unique approach to knowledge distillation

4 min readJan 4, 2021

--

CODIR: Train smaller, faster NLP models

--

Rohit Pillai
in
Towards Data Science

SpanFact: Fix your factually incorrect summaries

A suite of models that fixes inaccuracies in summaries of large text

5 min readDec 16, 2020

--

SpanFact: Fix your factually incorrect summaries

--

Rohit Pillai
in
Towards Data Science

HERO: You’ll never have to watch long videos again

A model that understands a video and its subtitles to help you speed through it

4 min readSep 17, 2020

--

HERO: You’ll never have to watch long videos again

--

Rohit Pillai
in
The Startup

FILTER: Understand Foreign Languages Better

A technique that promotes a deeper understanding and linkage between the same text in different languages

5 min readSep 16, 2020

--

FILTER: Understand Foreign Languages Better

--

Rohit Pillai
in
Towards Data Science

DiscoBERT: A BERT that shortens your reading time

A Discourse Aware BERT that improves summary generation

4 min readAug 22, 2020

--

DiscoBERT: A BERT that shortens your reading time

--

Rohit Pillai
in
The Startup

Distill-BERT: Using BERT for Smarter Text Generation

Generate better stories, essays, answers and more using BERT’s knowledge

5 min readAug 7, 2020

--

Distill-BERT: Using BERT for Smarter Text Generation

--

Rohit Pillai
in
Towards Data Science

VILLA: A Generic Adversarial Training technique for Vision-and-Language

The first multimodal adversarial training technique for image and text models

4 min readJul 13, 2020

--

VILLA: A Generic Adversarial Training technique for Vision-and-Language

--

Rohit Pillai
in
The Startup

BachGAN: Let your imagination run free

Create real images with just boxes!

6 min readMay 15, 2020

--

BachGAN: Let your imagination run free

--

Rohit Pillai
in
The Startup

VIOLIN: Do you really understand videos?

A new task and dataset that requires multi-modal understanding of videos and text

4 min readMay 3, 2020

--

VIOLIN: Do you really understand videos?

--

Rohit Pillai
in
Towards Data Science

UNITER

Learning a joint representation of image and text that everything can use

6 min readMar 5, 2020

--

1

UNITER

--

1

Rohit Pillai

Rohit Pillai

I’m an engineer at Microsoft Dynamics 365 AI Research and I’ll post our new NLP, CV and Multimodal research . Check out https://medium.com/@rohit.rameshp

Help
Status
About
Careers
Blog
Privacy
Terms
Text to speech
Teams