Dhruv Verma

SWE @microsoft, Grad student @stonybrooku

[Dhruv Verma @dhr_verma]

@Zipcar still waiting for your awful customer service to get back to me with the refund.

[Dhruv Verma @dhr_verma]

To add to this, this is the second time in 3 months my zipcar has broken down for unknown reasons. Please invest in maintaining your rental cars.

[Dhruv Verma @dhr_verma]

@Zipcar your service is absolutely terrible. My rental broke down at 7am at Rainier National Park with no service at all!

[Dhruv Verma @dhr_verma]

And still no sign of the refund for my trip. Truly terrible service. I’ll be unsubscribing today.

[Adithya V Ganesan @adigan8]

1/ 🚨 Our latest research work, soon to be presented at @wassa_ws during #ACL2023NLP #NLProc, reveals surprising insights about #GPT3’s ability to estimate the big 5 personality traits. Here are some intriguing findings:

[Dhruv Verma @dhr_verma]

About to present our work at the Frontenac Ballroom from 11 at @!

[Dhruv Verma @dhr_verma]

Excited to present today during the first poster session of @aclmeeting from 11am to 12:30pm! Defn drop by to a chat about our work! #NLProc #ACL2023

[Tech At Bloomberg @TechAtBloomberg]

The poster for "Evaluating Paraphrastic Robustness in Textual Entailment Models," joint work by @dhr_verma. @lal_yash, our @SinhaShreyashee, @azpoliak & @ben_vandurme, is being presented during Poster Session 1 (11 AM EDT) at @aclmeeting
#ACL2023NLP #NLProc

[ACL 2026 @aclmeeting]

📢 Exciting news! The #ACL2023NLP conference kicks off today with six interesting tutorials. Here's what's on the agenda before lunch: (1/5)

[Dhruv Verma @dhr_verma]

At @aclmeeting! Can’t wait for the keynote by @geoffreyhinton !

#ACL2023 #NLProc

[Dhruv Verma @dhr_verma]

Excited to present today during the first poster session of @aclmeeting from 11am to 12:30pm! Defn drop by to a chat about our work! #NLProc #ACL2023

[Niranjan @b_niranjan]

@stonybrooku and collaborators have an exciting lineup of papers in the @aclmeeting this year. Also a largish contingent of our students are attending. Check out their talks and posters.

Unfortunately I will taste fomo this year having just moved to Korea.

[Dhruv Verma @dhr_verma]

Excited to present today during the first poster session of @aclmeeting from 11am to 12:30pm! Defn drop by to a chat about our work! #NLProc #ACL2023

[Dhruv Verma @dhr_verma]

Current models are good at recognizing textual entailment (RTE). But if an RTE model has a sufficiently high capacity for reliable, robust inference necessary for NLU, then the model's predictions should be consistent across paraphrased examples. #NLProc (1/6)

[Dhruv Verma @dhr_verma]

Heading to Toronto today to attend @aclmeeting. Going to be presenting our work tomorrow morning during the first poster session! Do drop by to learn more about our work! #NLProc #ACL2023

[Dhruv Verma @dhr_verma]

Current models are good at recognizing textual entailment (RTE). But if an RTE model has a sufficiently high capacity for reliable, robust inference necessary for NLU, then the model's predictions should be consistent across paraphrased examples. #NLProc (1/6)

[Dhruv Verma @dhr_verma]

Excited to present our work at @aclmeeting on July 10th during the first poster presentation session! #NLProc #ACL2023

[Dhruv Verma @dhr_verma]

Current models are good at recognizing textual entailment (RTE). But if an RTE model has a sufficiently high capacity for reliable, robust inference necessary for NLU, then the model's predictions should be consistent across paraphrased examples. #NLProc (1/6)

[Dhruv Verma @dhr_verma]

Excited to be at @aclmeeting between July 9th and July 12th. Do drop by to view our poster!

[Dhruv Verma @dhr_verma]

@lal_yash @SinhaShreyashee @azpoliak @ben_vandurme @aclmeeting The paper can be found at https://t.co/h31GhbUICi. The dataset can be found at https://t.co/TkZKdv491A. #ACL2023 #NLProc (6/6)

[Adam Poliak @azpoliak]

@dhr_verma will be presenting our work at @aclmeeting. Make sure to check it out!

[Dhruv Verma @dhr_verma]

@lal_yash @SinhaShreyashee @azpoliak @ben_vandurme @aclmeeting The paper can be found at https://t.co/h31GhbUICi. The dataset can be found at https://t.co/TkZKdv491A. #ACL2023 #NLProc (6/6)

[Dhruv Verma @dhr_verma]

Current models are good at recognizing textual entailment (RTE). But if an RTE model has a sufficiently high capacity for reliable, robust inference necessary for NLU, then the model's predictions should be consistent across paraphrased examples. #NLProc (1/6)

[Dhruv Verma @dhr_verma]

This lack of consistency implies that models have room to improve on paraphrased data. This is joint work w/ @lal_yash, @SinhaShreyashee, @azpoliak and @ben_vandurme. Excited to present at @aclmeeting #ACL2023 #NLProc (5/6)