OpenReview
.net
OpenReview
.net
Login
OpenReview
.net
Login
Lachlan Ewen MacDonald
Postdoc, University of Pennsylvania, University of Pennsylvania
Joined
May 2021
Names
Lachlan Ewen MacDonald
(Preferred)
,
Lachlan E. MacDonald
Emails
****@adelaide.edu.au
(Confirmed)
,
****@protonmail.com
(Confirmed)
Personal Links
Homepage
Google Scholar
DBLP
Career & Education History
Postdoc
University of Pennsylvania, University of Pennsylvania
(upenn.edu)
2024
–
Present
Postdoc
Johns Hopkins University
(jhu.edu)
2023
–
2024
Postdoc
University of Adelaide
(adelaide.edu.au)
2021
–
2023
Advisors, Relations & Conflicts
Coworker
Hancheng Min
2023
–
2025
Coauthor
Ziqing Xu
2023
–
2025
Coworker
Salma Tarmoun
2023
–
2025
Postdoc Advisor
Rene Vidal
2023
–
2023
Coauthor
Jack Valmadre
2022
–
2023
Postdoc Advisor
Simon Lucey
2021
–
2023
Coauthor
Sameera Ramasinghe
2021
–
2023
Coauthor
Hemanth Saratchandran
2020
–
2023
Expertise
Deep learning
2021
–
2025
Theory
2021
–
2025
Optimisation
2021
–
2025
Optimization
2021
–
2025
Generalization
2021
–
2025
Statistical learning theory
2021
–
2025
Publications
Softmax-Induced Ill-Conditioning in Transformer Models
Salma Tarmoun
,
Lachlan Ewen MacDonald
,
Ziqing Xu
,
Hancheng Min
,
Rene Vidal
Submitted to ICLR 2026
Readers:
Everyone
LoRA Provably Reduces Forgetting and Enables Adapter Merging in Multiclass Linear Classification
Ziqing Xu
,
Hancheng Min
,
Lachlan Ewen MacDonald
,
Salma Tarmoun
,
Weijie J Su
,
Enrique Mallada
,
Rene Vidal
Submitted to ICLR 2026
Readers:
Everyone
Convergence Rates for Gradient Descent on the Edge of Stability for Overparametrised Least Squares
Lachlan Ewen MacDonald
,
Hancheng Min
,
Leandro Palma
,
Salma Tarmoun
,
Ziqing Xu
,
Rene Vidal
NeurIPS 2025 poster
Readers:
Everyone
Understanding the Learning Dynamics of LoRA: A Gradient Flow Perspective on Low-Rank Adaptation in Matrix Factorization
Ziqing Xu
,
Hancheng Min
,
Lachlan Ewen MacDonald
,
Jinqi Luo
,
Salma Tarmoun
,
Enrique Mallada
,
René Vidal
CoRR 2025
Readers:
Everyone
Understanding the Learning Dynamics of LoRA: A Gradient Flow Perspective on Low-Rank Adaptation in Matrix Factorization
Ziqing Xu
,
Hancheng Min
,
Lachlan Ewen MacDonald
,
Jinqi Luo
,
Salma Tarmoun
,
Enrique Mallada
,
Rene Vidal
AISTATS 2025 Poster
Readers:
Everyone
Gradient Descent and Attention Models: Challenges Posed by the Softmax Function
Salma Tarmoun
,
Lachlan Ewen MacDonald
,
Hancheng Min
,
Ziqing Xu
,
Rene Vidal
ICLR 2025 Conference Withdrawn Submission
Readers:
Everyone
D'OH: Decoder-Only Random Hypernetworks for Implicit Neural Representations
Cameron Gordon
,
Lachlan E. MacDonald
,
Hemanth Saratchandran
,
Simon Lucey
ACCV (7) 2024
Readers:
Everyone
D'OH: Decoder-Only random Hypernetworks for Implicit Neural Representations
Cameron Gordon
,
Lachlan Ewen MacDonald
,
Hemanth Saratchandran
,
Simon Lucey
CoRR 2024
Readers:
Everyone
On progressive sharpening, flat minima and generalisation
Lachlan Ewen MacDonald
,
Jack Valmadre
,
Simon Lucey
Submitted to ICLR 2024
Readers:
Everyone
On skip connections and normalisation layers in deep optimisation
Lachlan Ewen MacDonald
,
Jack Valmadre
,
Hemanth Saratchandran
,
Simon Lucey
NeurIPS 2023 poster
Readers:
Everyone
View all 28 publications
Co-Authors
Cameron Gordon
Chaoyang Wang
Enrique Mallada
Hancheng Min
Hemanth Saratchandran
Jack Valmadre
Jinqi Luo
Leandro Palma
László A. Jeni
Moshiur Farazi
Moshiur R Farazi
Moshiur R. Farazi
Rene Vidal
René Vidal
Salma Tarmoun
Sameera Ramasinghe
Shin-Fang Ch'ng
Shin-Fang Chng
Simon Lucey
Weijie J Su
Ziqing Xu