Sagar Shrestha

Sagar Shrestha Email and Phone Number

Research Intern, Multimodal Learning @ Samsung Research America (SRA)
Corvallis, OR, US
Sagar Shrestha's Location
Corvallis, Oregon, United States, United States
About Sagar Shrestha

I am a PhD student advised by Dr. Xiao Fu at Oregon State. I work on developing principled methods for unsupervised learning, multi-modal learning, and generative models.

Sagar Shrestha's Current Company Details
Samsung Research America (SRA)

Samsung Research America (Sra)

View
Research Intern, Multimodal Learning
Corvallis, OR, US
Sagar Shrestha Work Experience Details
  • Samsung Research America (Sra)
    Research Intern, Multimodal Learning
    Samsung Research America (Sra)
    Corvallis, Or, Us
  • Oregon State University
    Graduate Research Assistant
    Oregon State University Jan 2021 - Present
    Corvallis, Oregon, United States
    - Developed unified framework for image generation and translation, which enabled generating image (e.g., conditional GAN) and translating between domains (e.g., cat to dog) using the same model.- Proposed a provable method for extracting latent variables from multi-view unpaired data, which showed state-of-the-art performance for domain adaptation, multi-view retrieval, and single cell sequence alignment.- Developed first-of-its-kind identifiability guaranteed method for unsupervised… Show more - Developed unified framework for image generation and translation, which enabled generating image (e.g., conditional GAN) and translating between domains (e.g., cat to dog) using the same model.- Proposed a provable method for extracting latent variables from multi-view unpaired data, which showed state-of-the-art performance for domain adaptation, multi-view retrieval, and single cell sequence alignment.- Developed first-of-its-kind identifiability guaranteed method for unsupervised domain translation based on CycleGAN framework and its variants, which resolved the inherent content misaligned translation issue of existing methods.- Provided explanation for why common encoder based image to image translation such as StarGAN are successful and how to improve them.- Developed novel communication efficient federated learning for multiview/multimodal data, which enabled privacy preserving distributed training of deep GCCA and reduced the communication overhead by more than 90% with essentially no loss in performance.- Proposed a imitation learning method based on graph neural network policy for speeding up branch and bound algorithm for antenna selection and beamforming problem, which exhibited strong generalization to large unseen problem sizes.- Proposed a novel method based on generative model (GAN), tensor decomposition, non-negative matrix factorization, Gaussian dithering, and maximum likelihood objective to solve the (quantized) spectrum cartography problem, which resulted in state-of-the-art performance in various challenging scenarios (e.g., heavy shadowing, large number of emitters, etc). Show less
  • Paaila Technology
    Co-Founder, Robotics And Ai Engineer
    Paaila Technology Oct 2016 - Nov 2020
    Kathmandu, Bāgmatī, Nepal
    - Designed and built food Delivery robot with autonomous navigation, which resulted in Nepal's first fully automated robotic restaurant in Kathmandu.- Designed and build Customer Service Chatbot Agent, which was deployed in multiple banks' websites in Nepal.

Sagar Shrestha Education Details

Frequently Asked Questions about Sagar Shrestha

What company does Sagar Shrestha work for?

Sagar Shrestha works for Samsung Research America (Sra)

What is Sagar Shrestha's role at the current company?

Sagar Shrestha's current role is Research Intern, Multimodal Learning.

What schools did Sagar Shrestha attend?

Sagar Shrestha attended Oregon State University, Oregon State University, Institute Of Engineering, Pulchowk Campus.

Not the Sagar Shrestha you were looking for?

Free Chrome Extension

Find emails, phones & company data instantly

Find verified emails from LinkedIn profiles
Get direct phone numbers & mobile contacts
Access company data & employee information
Works directly on LinkedIn - no copy/paste needed
Get Chrome Extension - Free

Download 750 million emails and 100 million phone numbers

Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.