Xiaodan Song
9 papers · 2020–2022 · 5 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+3 more ↓ Show less ↑
π£ Hot Topic Early Bird π Interdisciplinary Bridge π§ Keyword Pioneer π Conference Polyglot (5) π Cross-Pollinator (11)
π
Renaissance Researcher
(5)
πΊοΈ
Taxonomy Completionist
(17)
β‘
Prolific Year
(6)
Conferences
ECCV (3)
ACL (2)
ICLR (2)
CVPR (1)
ICML (1)
Top co-authors
Keywords
model compression
(3)
transfer learning
(2)
knowledge distillation
(2)
knowledge transfer
(1)
natural language inference
(1)
question answering
(1)
neural architecture search
(1)
deep learning
(1)
neural network optimization
(1)
computational efficiency
(1)
multi-scale feature
(1)
mobile deployment
(1)
model optimization
(1)
backbone model
(1)
masked language modeling
(1)
pretrained language model
(1)
encoder-decoder architecture
(1)
network initialization
(1)
feature pyramid
(1)
deep thin network
(1)
Papers
Token Dropping for Efficient BERT Pretraining
ACL 2022
Auto-scaling Vision Transformers without Training
ICLR 2022
A Simple Single-Scale Vision Transformer for Object Detection and Instance Segmentation
ECCV 2022
Efficient Scale-Permuted Backbone with Learned Resource Distribution
ECCV 2020
Large Batch Optimization for Deep Learning: Training BERT in 76 minutes
ICLR 2020
MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices
ACL 2020
Go Wide, Then Narrow: Efficient Training of Deep Thin Networks
ICML 2020
SpineNet: Learning Scale-Permuted Backbone for Recognition and Localization
CVPR 2020
BigNAS: Scaling Up Neural Architecture Search with Big Single-Stage Models
ECCV 2020