Kanishka Misra

Kanishka Misra

PhD Student

Purdue University

I am a third year PhD Student at Purdue University, where I work on Natural Language Understanding with Julia Taylor Rayz at the AKRaNLU Lab. I am particularly interested in analysing neural network representations of language from the perspective of human sentence processing.

In 2018, I was fortunate to be funded by the Purdue Research Foundation Fellowship. I currently teach three sections (~75 students) of database fundamentals to sophomore level undergraduates.

My email is [my-first-name] @ purdue [dot] edu.[why is it like that?]


  • Natural Language Processing
  • Lexical Semantics
  • Sentence Processing in Humans


  • Julia Rayz, Purdue (Chair)
  • Allyson Ettinger, UChicago
  • Tatiana Ringenberg, Purdue
  • Ananya Sheth, Purdue


  • PhD Student, Current

    Purdue University

  • BS in Computer Information Technology, 2018

    Purdue University

Recent News


  • September 2020: Paper with Allyson and Julia accepted for publication in Findings of EMNLP ! Preprint and Code coming soon!

  • September 2020: I was featured by the Purdue Graduate School on their social media! [Link].

  • August 2020: Congrats to my undergraduate advisees Qingyuan “Carol” Hu and Yi Zhang for their paper with Julia and me being accepted to ICCI*CC 2020!

  • July 2020: Presented a Poster with Allyson Ettinger and Julia at the first virtual CogSci!

  • May 2020: Paper with Julia accepted at NAFIPS'20!

  • May 2020: Poster with Allyson Ettinger and Julia accepted at CogSci 2020!

Recent Publications

Quickly discover relevant content by filtering publications.

Exploring Lexical Relations in BERT using Semantic Priming (Forthcoming)

Using semantic priming to investigate how BERT utilizes lexical relations to inform word probabilities in context.

An Approximate Perspective on Word Prediction in Context: Ontological Semantics meets BERT (Forthcoming)

Analyzing BERT’s word prediction in context from the lens of Ontological Semantics

A Sentiment Based Non-Factoid Question-Answering Framework

Adding sentiment based features to non-factoid question answering

Recent Posts

Implementing a Neural Network from scratch in R using rrays

Summary In this post, I want to briefly demonstrate how one can implement a simple Neural Network architecture in R using a new package for array/tensor computation called rray along with the R6 object oriented programming system which will be used to modularize the various elements requried for anyone to build a neural network architecture.

My first few open source contributions: Authorship Attribution of the Federalist Papers

Background During the last semester of my undergraduate education at Purdue, I was engaged in a research project that analyzed conversation between two participants, and delivered some insight regarding the two participants’ future interaction(this will be expanded further in a blog post maybe).

Population growth and Doubling times with tidyverse

Roses are red, violets are blue This is a forced rhyme, here’s blog post two! Background Ever since I worked on data about populations at my internship at Perscio, a healthcare data analysis firm in Indianapolis, as well as worked with a Professor of Demography and Social Policy on a paper about demographic data, I have gained interest in population problems - mostly through readings.

Attitudes of employees towards mental health in the tech workplace

Understanding and accepting mental health as an issue at the workplace has become ever so crucial in recent times. Mental illnesses like depression and anxiety can have a significant economic impact; the estimated cost to the global economy is US$ 1 trillion per year in lost productivity (source).

The First Post

Why? I feel no other content can do enough justice to a blog’s first post than the one describing why the author would take out time from his/her daily routine and write about stuff.