“sybill.ai is going to be huge. Nishit Asnani and his co-founders are at the cutting edge. Not only was I blown away by their AI product, I actually immediately understood exactly how it could complement me as a salesperson in a tremendous fashion. In the age of flashy gadgets that talk a lot of hype and don't deliver, sybill.ai offers something truly remarkable you don't want to pass up.”
About
it's quick, helpful, and patient.
It keeps me…
Activity
-
Hi everyone! Rajalakshmi Venkatesh here, Head of Talent & Internal Operations at HockeyStack. I built teams from scratch at Meta, Slack, and…
Hi everyone! Rajalakshmi Venkatesh here, Head of Talent & Internal Operations at HockeyStack. I built teams from scratch at Meta, Slack, and…
Liked by Nishit Asnani
-
I interview almost every HockeyStack customer as soon as they become a customer, and I always hear the same thing: Your gift made you stand out…
I interview almost every HockeyStack customer as soon as they become a customer, and I always hear the same thing: Your gift made you stand out…
Liked by Nishit Asnani
-
When I say this... I truly believe it... Clay's Waterfall Enrichment will become everyone's Email Finder/Mobile Finder. Their commitment to…
When I say this... I truly believe it... Clay's Waterfall Enrichment will become everyone's Email Finder/Mobile Finder. Their commitment to…
Liked by Nishit Asnani
Experience & Education
Publications
-
CheXaid: deep learning assistance for physician diagnosis of tuberculosis using chest x-rays in patients with HIV
NPJ Digital Medicine (Nature Partner Journals)
Tuberculosis (TB) is the leading cause of preventable death in HIV-positive patients, and yet often remains undiagnosed and untreated. Chest x-ray is often used to assist in diagnosis, yet this presents additional challenges due to atypical radiographic presentation and radiologist shortages in regions where co-infection is most common. We developed a deep learning algorithm to diagnose TB using clinical information and chest x-ray images from 677 HIV-positive patients with suspected TB from…
Tuberculosis (TB) is the leading cause of preventable death in HIV-positive patients, and yet often remains undiagnosed and untreated. Chest x-ray is often used to assist in diagnosis, yet this presents additional challenges due to atypical radiographic presentation and radiologist shortages in regions where co-infection is most common. We developed a deep learning algorithm to diagnose TB using clinical information and chest x-ray images from 677 HIV-positive patients with suspected TB from two hospitals in South Africa. We then sought to determine whether the algorithm could assist clinicians in the diagnosis of TB in HIV-positive patients as a web-based diagnostic assistant. Use of the algorithm resulted in a modest but statistically significant improvement in clinician accuracy (p = 0.002), increasing the mean clinician accuracy from 0.60 (95% CI 0.57, 0.63) without assistance to 0.65 (95% CI 0.60, 0.70) with assistance. However, the accuracy of assisted clinicians was significantly lower (p < 0.001) than that of the stand-alone algorithm, which had an accuracy of 0.79 (95% CI 0.77, 0.82) on the same unseen test cases. These results suggest that deep learning assistance may improve clinician accuracy in TB diagnosis using chest x-rays, which would be valuable in settings with a high burden of HIV/TB co-infection. Moreover, the high accuracy of the stand-alone algorithm suggests a potential value particularly in settings with a scarcity of radiological expertise.
-
Disambiguating Sentiment: An Ensemble of Humour, Sarcasm, and Hate Speech Features for Sentiment Classification
Workshop on Noisy User-generated Text (WNUT), Empirical Methods in Natural Language Processing (EMNLP), 2019
Due to the nature of online user reviews, sentiment analysis on such data requires a deep semantic understanding of the text. Many online reviews are sarcastic, humorous, or hateful. Signals from such language nuances may reinforce or completely alter the sentiment of a review as predicted by a machine learning model that attempts to detect sentiment alone. Thus, having a model that is explicitly aware of these features should help it perform better on reviews that are characterized by them. We…
Due to the nature of online user reviews, sentiment analysis on such data requires a deep semantic understanding of the text. Many online reviews are sarcastic, humorous, or hateful. Signals from such language nuances may reinforce or completely alter the sentiment of a review as predicted by a machine learning model that attempts to detect sentiment alone. Thus, having a model that is explicitly aware of these features should help it perform better on reviews that are characterized by them. We propose a composite two-step model that extracts features pertaining to sarcasm, humour, hate speech, as well as sentiment, in the first step, feeding them in conjunction to inform sentiment classification in the second step. We show that this multi-step approach leads to a better empirical performance for sentiment classification than a model that predicts sentiment alone. A qualitative analysis reveals that the conjunctive approach can better capture the nuances of sentiment as expressed in online reviews.
Projects
-
Wordsmith
-
Wordsmith is a platform for creative writers to share their work and receive feedback from other writers. Users can create private groups with people they want to share their work with or public groups where anyone can join, read and share. Wordsmith hosted a six-word story challenge with over 80 participants, and as such helps create a cohesive network of writers.
Other creatorsSee project -
Paraphrase Generation using Deep Generative Models
-
Developed a model to generate multiple paraphrases for a source sentence using a variational autoencoder with control variables in the latent space used to specify the length of the paraphrase required.
Achieved better BLEU scores than vanilla seq2seq models, generated sentences perform better than ground truth parphrases on human evaluation for MSCOCO dataset -
Explicit Exploration in Random Forests
-
Objective was to introduce an explicit exploration control knob in random forests for multi-label learning
Showed empirically that using UCB / Thompson Sampling over the features at each level of a random forest for splitting each node consistently beat other criteria like empirical mean (with ranking loss) for major binary classification datasets (spam, MNIST etc)
Languages
-
English
-
-
Hindi
-
Recommendations received
1 person has recommended Nishit
Join now to viewMore activity by Nishit
-
Joining the Early Stage team at Sequoia Capital 🌲 this summer! Super excited to be part of this legendary organization that has backed OpenAI…
Joining the Early Stage team at Sequoia Capital 🌲 this summer! Super excited to be part of this legendary organization that has backed OpenAI…
Liked by Nishit Asnani
-
99% of SaaS companies say they’re focused on their ideal customer profile (ICP). But only a *select few* could show me this chart 👇 What it means…
99% of SaaS companies say they’re focused on their ideal customer profile (ICP). But only a *select few* could show me this chart 👇 What it means…
Liked by Nishit Asnani
-
There are so many interesting papers about AI and every day it feels like there are a dozen more. There are four papers though that I would recommend…
There are so many interesting papers about AI and every day it feels like there are a dozen more. There are four papers though that I would recommend…
Liked by Nishit Asnani
Other similar profiles
-
Mehak Aggarwal
Connect -
Soumyarka Mondal
Connect -
Aarti Motiani
Hiring growth and marketing ninjas | Growth Mktg @ Sybill
Connect -
Ananya Jain
Connect -
Benjamin Sternsmith
Connect -
Ganeshan Malhotra
Sybill.ai | ML @ UCSD | Prev Quantiphi, UHamburg, IIITD, BITS-Pilani
Connect -
Nikki Gagnon
Connect -
Afroze Syed
MBA Candidate at UCLA Anderson | IIT Delhi
Connect -
Devlina Das
Connect -
Taylor Haren
Connect
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More