About

Leonid Karlinsky

Research Scientist, Senior Technical Staff Member

Who they work with

Categories

Artificial Intelligence Computer Vision Deep Learning Few-shot Learning

Leonid Karlinsky is a Principal Research Scientist (STSM) in the MIT-IBM lab. Prior to that Leonid led the AI Vision research group in the Multimedia department @ IBM Research AI. Leonid joined IBM Research in July 2015. Before joining IBM, he served as a research scientist in Applied Materials, Elbit, and FDNA. He is actively publishing, reviewing, and performing occasional chair duties at ECCV, ICCV, CVPR, ICLR, AAAI, WACV, and NeurIPS, and is serving as an IMVC steering committee member for the past 6 years. During his time at IBM, Leonid has co-authored over 25 research papers on his research is in the areas of augmented reality, medical applications, self-supervised, cross-domain, multi-modal, and few-shot learning. He received his PhD degree at the Weizmann Institute of Science, supervised by Prof. Shimon Ullman.

Publications with the MIT-IBM Watson AI Lab

ConStruct-VL: Data-Free Continual Structured VL Concepts Learning

Teaching Structured Vision & Language Concepts to Vision & Language Models

CODA-Prompt: COntinual Decomposed Attention-based Prompting for Rehearsal-Free Continual Learning

Learning to Grow Pretrained Models for Efficient Transformer Training

Multitask Prompt Tuning Enables Parameter-Efficient Transfer Learning

Contrastive Audio-Visual Masked Autoencoder

How Transferable are Video Representations Based on Synthetic Data?

Bringing Image Scene Structure to Video via Frame-Clip Consistency of Object Tokens

Computer Vision NeurIPS

FETA: Towards Specializing Foundation Models for Expert Task Applications

Self-supervised classification network

Task2Sim: Towards Effective Pre-training and Transfer from Synthetic Data

Unsupervised Domain Generalization by Learning a Bridge Across Domains

Dynamic Distillation Network for Cross-Domain Few-Shot Recognition with Unlabeled Data

A Broad Study on the Transferability of Visual Representations With Contrastive Learning

Detector-Free Weakly Supervised Grounding by Separation

StarNet: towards weakly supervised few-shot detection and explainable few-shot classification

AAAI Computer Vision

Fine-grained Angular Contrastive Learning with Coarse Labels

CVPR Computer Vision

AdaFuse: Adaptive Temporal Fusion Network for Efficient Action Recognition

ICLR Computer Vision

A Broader Study of Cross-Domain Few-Shot Learning

ECCV Computer Vision

OnlineAugment: Online Data Augmentation with Less Domain Knowledge

ECCV Computer Vision

TAFSSL: Task-Adaptive Feature Sub-Space Learning for few-shot classification

ECCV Computer Vision

AR-Net: Adaptive Frame Resolution for Efficient Action Recognition

ECCV Computer Vision

LaSO: Label-Set Operations networks for multi-label few-shot learning

Computer Vision Few-shot Learning

RepMet: Representative-based metric learning for classification and one-shot object detection

Computer Vision

Co-regularized Alignment for Unsupervised Domain Adaptation

Transfer Learning Deep Learning

Delta-encoder: an effective sample synthesis method for few-shot object recognition

Computer Vision Few-shot Learning