About

I am an MSc graduate at the IE&M Faculty of the Technion - Israel Institute of Technology, under the supervision of Prof. Roi Reichart.

My master's thesis presents the novel task of sarcasm interpretation, along with a publicly available dataset and a Machine Translation based sarcasm interpretation system with a sentiment analysis component (the Sarcasm SIGN).

My main interest in NLP is that grey area where machines fail to understand humans - non literal communication such as humor, irony and sarcasm. I enjoy defining new tasks and approaching them with different methods, and currently am most interested in neural sequence to sequence models.

I'm currently working as a Data Scientist and NLP researcher at Fiverr, building ML algorithms and Neural models for challenges in the online marketplace - soliciting for 5 stars reviews, sentiment analysis under complex settings, and much more.


P.S - I love teaching and giving talks and about Deep Learning and NLP , contact me if you're interested. :)


Publications

"Sarcasm SIGN: Interpreting Sarcasm with Monolingual Machine Translation." Lotem Peled and Roi Reichart. ACL 2017 (long paper) [arXiv preprint (arxiv:1704.06836): pdf] .


Lectures


Media \ Buzz

  • Fortune tech - "Getting Computers to Understand Sarcasm is Easy. Not!"
  • Business Cloud - "How Seinfeld is helping AI understand your sarcasm"
  • Jerusalem Post - "Technion system interprets Twitter sarcasm"
  • Technion Magazine - "מדעי הסרקזם"
  • Ynet - "סטודנטית בטכניון פיתחה מערכת המפרשת סרקזם בטוויטר"
  • Ma'ariv - "סטודנטית פיתחה מערכת חדשה שתזהה את הציניות שלכם"


Contact

lotemi.peled@gmail.com || LinkedIn