Sathappan
@sathappanspm
Followers
92
Following
1K
Media
0
Statuses
86
PhD candidate in Data mining and ML. Discovery Analytics Center, Virginia Tech
Joined September 2010
Piling on to the pile-on (sorry - it's always easy to criticize 😛), here's a rant about benchmarks for LLMs that are used to back claims of "stronger" or "better" models. Let's start with a tour through GPT-3's Appendix G... 1/8
The thread is mostly skeptical that any of this work will extrapolate in practice, but that was why we trained chinchilla 70B. It's not a perfect comparison, but trained on the same arch, codebase, dataset as gopher 280B for the same compute --- and it was a lot better 2/
5
11
120
Detecting Irregular Network Activity with Adversarial Learning and Expert Feedback https://t.co/bmCvGBBBWF by Gopikrishna Rathinavel et al. including @nikhilm_1, @profnaren
#DeepLearning #AnomalyDetection
deepai.org
10/01/22 - Anomaly detection is a ubiquitous and challenging task relevant across many disciplines. With the vital role communication network...
1
6
11
I'm excited to share my paper 'Framing Algorithmic Recourse for Anomaly Detection' with @profnaren @niustrategy will be presented at ACM SigKDD 2022, Research Track. If you're at #KDD2022 check out our talk & poster on 8/16! [ https://t.co/Klkw7e95zT]
@kdd_news
arxiv.org
The problem of algorithmic recourse has been explored for supervised machine learning models, to provide more interpretable, transparent and robust outcomes from decision support systems. An...
2
3
11
Hearty Congratulations to Dr Nikhil @nikhilm_1 and best wishes for his next stint as Assistant professor at stevens Institute. I am sure he will be a great mentor for his students..
Congratulations to @nikhilm_1 on a successful PhD defense of his work on using domain knowledge for ML under data paucity and noise! Thanks to a very supportive committee of @anujkarpatne (co-advisor), @StefanoErmon, Chang-Tien Lu, & Danesh Tafti, from 3 time zones!
1
0
3
Congratulations Dr Datta!!
Congratulations to Debanjan Datta (@devDdata ) who successfully defended his dissertation on Discovery and Analysis of Suspicious Trade Records! Many thanks to his committee of Profs. Chris North, @NElmqvist, Chang-Tien Lu, and Chandan Reddy.
0
0
1
Congratulations Dr. Subhodip @dataquidnunc !!
Subhodip Biswas @dataquidnunc becomes Dr. Subhodip Biswas today! He successfully defended his PhD work on algorithms for school redistricting to his committee of Chang-Tien Lu, Chris North, @anujkarpatne, and Ian Davidson (UC Davis)! In addition to his algorithmic research,
1
0
4
Unveiling website for our @cyber_cci supported synthetic network traffic generation project - work by @ShengzheXu & work-of-art site design by Nathan Self (@tripswithtires), w/ code for generating massive netflow data records. Send us comments & feedback! https://t.co/E9jATSV634
sanghani-projects.cs.vt.edu
Read about recent work by the Sanghani Center on synthetic network traffic generation.
0
3
9
Highly innovative, helpful and a necessary product in medical diagnositics...a huge congratulations to my friend and founder @Geethanjalirad and all members of @AdiuvoDiag
Our pitch for the MedTech Innovator Asia Pacific Best Video Competition 2021! https://t.co/UHyHtQYDUP Like and contribute towards fighting the epidemic of Chronic Skin and Soft Tissue Infections!!
2
0
1
Very informative discussion about : COVID-19 Aftermath - Lessons for Data Science at #SIAMSDM21 Saturday (05/01). The diverse panel comprised: Dr. Hamsa Batani (@hamsabastani) , Dr. Rumi Chunara (@rumichunara), Dr. Madhav Marathe & Dr. Alex Vespignani (@alexvespi) --- 1/n
1
2
9
Please join us for presentation of our paper: PhyNet: Physics Guided Neural Networks for Particle Drag Force Prediction ( https://t.co/9eRCMepLQx) at #SIAMSDM21 at 10:30 AM by Nikhil Muralidhar et al. at the CP12 Time Series I session at #SIAMSDM21 @SIAMDataMining – 1/3
epubs.siam.org
Abstract Physics-based simulations are often used to model and understand complex physical systems in domains like fluid dynamics. Such simulations although used frequently, often suffer from...
1
1
3
Please join @realhuawei and me at the #SIAMSDM21 session we are chairing ( https://t.co/qdFS3Shhb9) on 04/30 from 10:30 AM - 12:30 PM. Interesting papers ranging from time series anomaly detection, segmentation & physics guidance in temporal contexts will be presented by authors.
1
3
7
Congrats to my PhD student Nikhil @nikhilm_1 on receiving a 2021 Applied ML Summer Research Fellowship to work with researchers at @LosAlamosNatLab on physics-informed ML for modeling adsorption equilibria in fluid mixtures! @anujkarpatne @VT_CS @SanghaniCtrVT @UrbComp
1
3
28
Patrick Butler, senior research associate, and @nikhilm_1, Ph.D. student @VT_CS @UrbComp, explain how @SanghaniCtrVT uses #BigData to forecast #COVID19 spread. More unified reporting, standardization of data collection could help future studies. #TechOnTapVT #VTInnovationCampus
.@SanghaniCtrVT is using #BigData to predict #COVID19 spread, model interventions, and help us avoid future pandemics or outbreaks #VTInnovationCampus
0
4
9
Detecting Anomalies Through Contrast in Heterogeneous Data https://t.co/ODy0q0cdj6 by Debanjan Datta et al. including @profnaren
#Autoencoder #AnomalyDetection
deepai.org
04/02/21 - Detecting anomalies has been a fundamental approach in detecting potentially fraudulent activities. Tasked with detection of illeg...
0
5
12
Ph.D. student @nikhilm_1 @SanghaniCtrVT talks about the model that the DeepOutbreak team developed to win two COVID-19 forecasting challenges. @profnaren @anika_anikat @GTCSE @UIowaCS
https://t.co/OiODWVBuds
0
2
9
Congrats to Raihan @Raihan0155 for a successful PhD thesis defense of his work on detecting and mitigating rumors in social media! Many thanks to his committee: Chang-Tien Lu, @liuhuan, @chandankreddy, and @badityap. We also answered the perennial question:
7
2
28
Gopikrishna Rathinavel’s interest in machine learning stemmed from courses in biotechnology. Pioneering research and opportunities for cross collaboration drew this master’s degree student to @VT_DAC @VT_CS. Student Spotlight: https://t.co/uWPnEV8DTz
@rgopikrishna_
@profnaren
0
4
10
Please join us this Friday April at the 24 at the ELLIS Geometric Deep Learning virtual workshop. We have an amazing lineup of speakers and the invited talks will all be streamed. Thanks to Thomas Kipf and Erik Bekkers for the organization. https://t.co/AawEaCw4la
1
50
228
When we compare the sizes of things, we tend to do it by comparing the closest edges. Which is why stuff like this happens!
3
30
95