PhD in data science from @InfAtEd. Likes Bayesian optimisation. Learning about rain in Bergen at NORCE and the Bjerknes centre. Opinions are my own.
Joined August 2020
- Tweets74
- Following181
- Followers158
- Likes194
Sigrid Passano Hellan retweeted
Now thesis is in — I’m on the job market! Looking for RS roles for product and/or research.
My PhD focused on cross-lingual transfer + structure prediction but also interested in analysis of optimisation, training dynamics and evaluating generation 🧵
Sigrid Passano Hellan retweeted
📢 Looking forward to presenting arxiv.org/abs/2311.11108 at #NeurIPS2023! If you think that model interpretability and accuracy for concept-based models are mutually exclusive, then come and check our poster out on 14th Thursday #1516
Sigrid Passano Hellan retweeted
I have funding for (3.5 yr) PhD positions at Edinburgh in #ML #AI.
Interested in safety & trustworthiness of systems that combine traditional software components, symbolic logics, and machine learning systems?
Apply at ed.ac.uk/informatics/postgra…
Join @InfAtEd -> @EdinburghUni
Sigrid Passano Hellan retweeted
Now accepted to NeurIPS 2023!
Thanks to the reviewers for helpful feedback and amazing collaborators @mueller_mp, @david_rolnick and Matthias Hein.
‼️ New paper: Normalization Layers Are All That Sharpness-Aware Minimization Needs ‼️
arxiv.org/abs/2306.04226
We show that applying SAM only to the normalization layers of a network (SAM-ON) enhances performance compared to applying it to the full network.
Sigrid Passano Hellan retweeted
I'm very excited to be in Seoul to give a talk and present a poster about our CVPR'23 Meta Omnium paper at the Hyundai Vision Conference! Feel free to take a look at the poster and also a few photos from this great city! Meta Omnium website: edi-meta-learning.github.io/…
Sigrid Passano Hellan retweeted
📢📢📢 Job Opportunity!
Join Dr Elliot J. Crowley and the Bayesian and Neural Systems Group as a postdoc in NAS and AutoML at the University of Edinburgh.
24-month funded position starting Autumn 2023. PhD in ML or related field required. Apply now: elxw.fa.em3.oraclecloud.com/…
Sigrid Passano Hellan retweeted
How can an RL agent adapt to new tasks without finetuning? In a nutshell, it can repeatedly update and sample from beliefs about the current task. In my @DeepMind internship work, we approximate these beliefs with a transformer and small graphs. arxiv.org/abs/2302.04250
Sigrid Passano Hellan retweeted
I’m very excited to come to #ICLR’23 in person! Together with @610v4nn1__ we’ll be presenting a poster for our paper “PASHA: Efficient HPO and NAS with Progressive Resource Allocation” tomorrow 1/5 at 16:30-18:30 CAT (MH1-2-3-4 #87). Looking forward to meeting some of you there!
Sigrid Passano Hellan retweeted
I’m excited to share that our PASHA paper has been accepted to ICLR’23! PASHA allows you to quickly tune the hyperparameters and is already available in Syne Tune github.com/awslabs/syne-tune…
If you need to tune the hyperparameters (or the architecture structure) of you ML algo but you don't have much time, take a look at our ICLR2023 paper: openreview.net/forum?id=syfg…
@OBohdal
@lukas_balles
@cedapprox
Sigrid Passano Hellan retweeted
Call for reviewers for our #ICLR2023 workshop on Mathematical and Empirical Understanding of Foundation Models.
Fill out this form if you are interested
sites.google.com/view/me-fom…
and we will aim to get back to you asap.
Paper deadline: 3 Feb
Tentative reviewing period: 10-24 Feb.
Sigrid Passano Hellan retweeted
ML & Climate Change social today @NeurIPSConf, 6pm, room 392. Roundtable topics: ML & Scientific Discovery, Working with Real-World Data, ML for Earth Monitoring, Biodiversity Monitoring, most impactful areas of research, and how to find jobs and interdisciplinary collaborations
Sigrid Passano Hellan retweeted
I’m excited to present “Feed-Forward Source-Free Domain Adaptation via Class Prototypes” at #ECCV OOD-CV workshop today! You can learn more about our new simple and fast method for adapting pre-trained models to new domains at
ood-cv.org/camera-ready/18/F…
Sigrid Passano Hellan retweeted
Curious about how to make HPO scalable to large datasets? You can learn more in our recent paper “PASHA: Efficient HPO with Progressive Resource Allocation” that I’ll present at the workshop track of the first #AutoML conference today. Full version at arxiv.org/abs/2207.06940.
Sigrid Passano Hellan retweeted
Very excited to talk about my @DeepMind internship work with @kroscoo and @TomasKocisky!
If you are attending #ICML2022 come to our spotlight talk/poster on Thursday to chat about entity-aware long narrative generation!
spotlight@10:45am, poster@6pm
icml.cc/virtual/2022/spotlig…
Headed to Baltimore for #ICML2022 next week? Our team will be presenting 30 papers, exploring themes like effective RL, language models, algorithmic reasoning, and much more! Read more in today’s blog: dpmd.ai/ICML22-Blog
Sigrid Passano Hellan retweeted
"Zero Shot Cross-Lingual Semantic Parsing" will be presented in person at @aclmeeting . I'll be at Poster Session 4: Semantics in the Forum at 11AM 24/5 and Virtual Poster Session 4 at 25/5
Anthology: aclanthology.org/2022.acl-lo…
🧵 #NLProc #ACL2022 #acl2022nlp
Sigrid Passano Hellan retweeted
The #ANC Seminar series in #ML #AI & #neuroscience reboots tomorrow with
👉Asa Cooper Stickland (@AsaCoopStick)
"Consistency of Neural Networks is Related to Loss Landscape Geometry"
👉Sigrid Hellan (@sighellan)
"Bayesian Optimisation for Active Monitoring of Air Pollution"
Sigrid Passano Hellan retweeted
If you're interested in differentiating over long optimization processes for meta-learning/HPO please come check out our #NeurIPS poster today 16:30 UTC!
Sigrid Passano Hellan retweeted
Would you like to learn how to make meta-learning more scalable? We’ll be presenting EvoGrad at Poster Session 1 at #NeurIPS today - starting from 16:30 UTC. Joint work with Yongxin Yang and Timothy Hospedales.
Sigrid Passano Hellan retweeted
Congratulations to @EdinburghNLP members @vernadankers and Kate McCurdy and their co-authors for Best Paper at @conll_conf !
Sigrid Passano Hellan retweeted
How Well Do Self-Supervised Models Transfer? Find out in our #CVPR presentation! Come chat with us in session 4 on Tuesday (11am EDT). Work done with @henrygouk & @tmh31. Paper: arxiv.org/abs/2011.13377. Video: youtu.be/prZiTphuzO0. #CVPR2021 (1/n)
