AFLoRA: Adaptive Freezing of Low Rank Adaptation in Parameter Efficient Fine-Tuning of Large Models
A novel Parameter-Efficient Fine-Tuning (PEFT) method that achieved SOTA performance while yeilding up to fewer average trainable parameters, runtime, and FLOPs.
ACL 2024, March 2024
03/2024
LMUFormer: Low Latency Yet Powerful Spiking Legendre Memory Units
A spiking model for long-sequential data that demonstrates comparable performance to Transformer and RNN models while necessitating a remarkable decrement in parameters and FLOPs.
ICLR 2024, January 2024
01/2024
Induced Generative Adversarial Particle Transformers
A novel GAPT model that offers linear time complexity and captures intricate jet substructure, surpassing MPGAN in many metrics.
Workshop at the 37th conference on NeurIPS, December 2023
12/2023
Spiking Neural Networks with Dynamic Time Steps for Vision Transformers
A novel training framework that dynamically allocates the number of time steps to each ViT module depending on a trainable score assigned to each timestep.
November 2023
11/2023
Evaluating Generative Models in High Energy Physics
The first systematic review and investigation into evaluation metrics and their sensitivity to failure modes of generative models, using the framework of two-sample goodness-of-fit testing, and their relevance and viability for HEP.
Physical Review D, 2022, November 2022
11/2022
