OpenReview
.net
OpenReview
.net
Login
OpenReview
.net
Login
Puneet K. Dokania
Principal Researcher, Five AI Limited, Bosch
Senior Researcher, Department of Engineering Science, University of Oxford
Joined
March 2018
Names
Puneet K. Dokania
(Preferred)
,
Puneet Dokania
,
Puneet Kumar Dokania
Emails
****@robots.ox.ac.uk
(Confirmed)
,
****@gmail.com
(Confirmed)
Personal Links
Homepage
Google Scholar
DBLP
Semantic Scholar
Career & Education History
Principal Researcher
Five AI Limited,
Bosch
(bosch.com)
2020
–
Present
Senior Researcher
Department of Engineering Science,
University of Oxford
(oxford.ac.uk)
2019
–
Present
Postdoc
Department of Engineering Science,
University of Oxford
(oxford.ac.uk)
2016
–
2019
PhD student
Computer Science and Applied Mathematics,
INRIA and CentraleSupelec
(inria.fr)
2012
–
2016
Advisors, Relations & Conflicts
Advisor
Prof. Philip Torr
2016
–
Present
PhD Advisor
M. Pawan Kumar
2012
–
2016
PhD Advisor
Nikos Paragios
2012
–
2016
Expertise
Computer Vision
Present
Deep Learning
Present
Reliability of AI models
Present
Continual Learning
Present
Publications
Beyond Mean Shifts: Predicting Distributional Responses to Unseen Genetic Perturbations
Kalyan Ramakrishnan
,
Jonathan G. Hedley
,
Sisi Qu
,
Puneet K. Dokania
,
Philip Torr
,
Cesar A. Prada-Medina
,
Julien Fauqueur
,
Kaspar Märtens
MLGenX 2026 Poster
Readers:
Everyone
Foundation Models for Trajectory Planning in Autonomous Driving: A Review of Progress and Open Challenges
Kemal Oksuz
,
Alexandru Buburuzan
,
Anthony Knittel
,
Yuhan Yao
,
Puneet K. Dokania
Accepted by TMLR
Readers:
Everyone
Mixture of Experts Made Intrinsically Interpretable
Xingyi Yang
,
Constantin Venhoff
,
Ashkan Khakzar
,
Christian Schroeder de Witt
,
Puneet K. Dokania
,
Adel Bibi
,
Philip Torr
ICML 2025 poster
Readers:
Everyone
MoCaE: Mixture of Calibrated Experts Significantly Improves Object Detection
Kemal Oksuz
,
Selim Kuzucu
,
Tom Joy
,
Puneet K. Dokania
Accepted by TMLR
Readers:
Everyone
What Makes and Breaks Safety Fine-tuning? A Mechanistic Study
Samyak Jain
,
Ekdeep Singh Lubana
,
Kemal Oksuz
,
Tom Joy
,
Philip Torr
,
Amartya Sanyal
,
Puneet K. Dokania
ICML 2024 MI Workshop Spotlight
Readers:
Everyone
What Makes and Breaks Safety Fine-tuning? A Mechanistic Study
Samyak Jain
,
Ekdeep Singh Lubana
,
Kemal Oksuz
,
Tom Joy
,
Philip Torr
,
Amartya Sanyal
,
Puneet K. Dokania
NeurIPS 2024 poster
Readers:
Everyone
RanDumb: Random Representations Outperform Online Continually Learned Representations
Ameya Prabhu
,
Shiven Sinha
,
Ponnurangam Kumaraguru
,
Philip Torr
,
Ozan Sener
,
Puneet K. Dokania
NeurIPS 2024 poster
Readers:
Everyone
Fine-tuning can cripple your foundation model; preserving features may be the solution
Jishnu Mukhoti
,
Yarin Gal
,
Philip Torr
,
Puneet K. Dokania
Accepted by TMLR
Readers:
Everyone
On Calibration of Object Detectors: Pitfalls, Evaluation and Baselines
Selim Kuzucu
,
Kemal Oksuz
,
Jonathan Sadeghi
,
Puneet K. Dokania
CoRR 2024
Readers:
Everyone
What Makes and Breaks Safety Fine-tuning? A Mechanistic Study
Samyak Jain
,
Ekdeep Singh Lubana
,
Kemal Oksuz
,
Tom Joy
,
Philip H. S. Torr
,
Amartya Sanyal
,
Puneet K. Dokania
CoRR 2024
Readers:
Everyone
View all 115 publications
Co-Authors
Adel Bibi
Adriana Romero-Soriano
Aivar Sootla
Albert Gordo
Ales Leonardis
Alexandru Buburuzan
Alexei A Efros
Alexei A. Efros
Amartya Sanyal
Ameya Prabhu
Ana Belen Espinosa-Gonzalez
Andrew Gambardella
Anthony Knittel
Anton Osokin
Ara Darzi
Arnab Ghosh
Arslan Chaudhry
Aseem Behl
Ashish Shah
Ashkan Khakzar
Atilim Günes Baydin
Bernard Ghanem
Blaine Rogers
Bor-Chun Chen
Botos Csaba
View all 121 co-authors