Deep Neural Network Compression
We create new compression algorithms to shrink the compute and memory footprint of deep neural networks.
We are rethinking the next generation of AI-centric computing systems, by developing DNN compression methods, efficient hardware architectures, and end-to-end AI systems optimization.
News
July 2025
Many collaborations have recently led to publications: Flashdepth at ICCV’25, Quamba2 at ICML’25, and Double Duty at FPL’25.
February 2025
Mohamed gave talks at Tower Research conference on synthetic software, and the LG AI Seminar, and Meta in NYC.
September 2024
FLIQS accepted to AutoML’24 and won best-paper award! Also, Kratos accepted to FPL’24.
August 2024
Students’s t-distributions, Generative NAS, and NAS Encodings accepted to ICML’24.
June 2024
Beyond Inference accepted to DAC’24 and NAS Latency Predictors accepted to MLSYS’24. Also, PQA accepted to TRETS/FCCM’24.
January 2024
Mohamed received the NSF CAREER Award to co-design efficient LLM hardware, software, and algorithms.
January 2024
Mohamed gave talks at Qualcomm Research, Yale University, and KAUST on efficient machine learning.
June 2023
Our group received an NSF Award to study fine-grained DNN sparsity.
May 2023
Multi-Predict accepted to AutoML’23 and BRAMAC accepted to FCCM’23.
February 2023
Zero-Cost Operation Scoring accepted to AAAI and our extended work on Logic Shrinkage has been accepted to TRETS.
December 2022
Mohamed gave talks at Zewail UST, Rutgers Efficient AI Seminar, Untether AI, and the FAI Summit.
October 2022
Adaptable Butterfly Accelerator accepted to MICRO’22 and BLOX accepted to NeurIPS’22 D&B Track.
October 2022
Our group received an Intel grant to study hardware-accelerated DNN inference.
September 2022
Mohamed gave a keynote at the International Symposium for Applied Reconfigurable Computing, and a talk at the AutoML seminar.
August 2022
Our group received a Meta Research Award in Networking for AI.
April 2022
Mohamed gave a talk at the Crossroads FPGA seminar.
March 2022
We received a TCS Research Award to study heterogeneous DNN computing.
January 2022
Logic Shrinkage paper accepted to FPGA’22 and nominated for best-paper award!
January 2022
Mohamed and Jordan joined the International Centre for Spatial Computational Learning.
January 2022
The Abdelfattah Research Group is formed at Cornell University, in the NYC Cornell Tech campus.
Research Themes
Efficient AI Systems
We analyze and optimize AI systems that host, shard, compress, and deploy large language models.
Hardware/software Codesign
We design new hardware architectures to accelerate efficient AI algorithms.
Reconfigurable Computing
We rethink reconfigurable computing architectures for low-latency custom AI computing.
▸ Our research is generously supported by amazing collaborators and sponsors, listed here.
People
Alumni
-
Marta Andronic
Visiting Researcher -
Angela Cui
Undergrad Researcher
⤷ PhD at Berkeley -
Anthony Fei
Undergrad Researcher -
Yifei Gao
MEng Researcher -
Yassine Ghannane
MS Student
⤷ PhD at U of Copenhagen -
Rebecca Helling
MEng Researcher -
Zhanqiu Hu
Undergrad Researcher
⤷ PhD at Cornell -
Rahul Jain
MEng Researcher
⤷ MIT Lincoln Lab -
Aditya Kotha
Undergrad Researcher
⤷ MS at USC -
Kristine Pham
Undergrad Researcher -
Junius Pun
Undergrad Researcher
⤷ Espressif Systems -
Buqing Xu
MEng Researcher -
Ussama Zahid
Visiting Researcher -
Allan Zhang
High School Researcher