Efimia Panagiotaki

38 posts

Efimia Panagiotaki banner
Efimia Panagiotaki

Efimia Panagiotaki

@panefimia

PhD Student @UniofOxford @oxfordrobots 🤖| @GoogleDeepMind Scholar | Sr ML Eng @Oxa_UA | Past @streetdrone @WilliamsRacing 🏎️, @ETH_en, @CVL_ETH, @ecentua

Oxford, England Katılım Eylül 2022
257 Takip Edilen117 Takipçiler
Efimia Panagiotaki retweetledi
Daniele De Martini
Daniele De Martini@d_d_martini·
Huge congratulations to @panefimia! 🎉 We're incredibly proud that her paper GraphSCENE is a finalist for the IEEE-RAS Outstanding Women in Robotics & Automation (WiRA) Paper Award @ #IROS2025! 🏅👏 She will also present it then at the event on Wed at noon.
English
1
1
1
74
Efimia Panagiotaki retweetledi
Daniele De Martini
Daniele De Martini@d_d_martini·
Introspection in Learned Semantic Scene Graph Localisation 🗺️ We probe how semantics affects semantic localization, prioritizing salient relations and distinctive landmarks for explainable registration. When: Fri, 16:25–16:35 • FAST workshop arxiv.org/pdf/2510.07053
English
1
1
1
52
Efimia Panagiotaki retweetledi
Daniele De Martini
Daniele De Martini@d_d_martini·
GraphSCENE: on-demand critical scenario generation for AVs 🚗 Temporal scene graphs, learned from real data and constrained by ontology, can generate novel scenarios in sim for testing AVs. When: Wed, 17:05–17:10 • WeDT17 efimiap.github.io/graphscene/
English
1
1
1
47
Efimia Panagiotaki retweetledi
Daniele De Martini
Daniele De Martini@d_d_martini·
The Oxford RobotCycle Project 🚴‍♂️🤖 A multimodal urban cycling dataset (range, visual, eye-gaze) + annotated maps, pointclouds, ontology, traffic — to study risk, attention, and infrastructure effects on VRU safety. When: Wed, 13:55-14:00 • WeBT17 ori-mrg.github.io/robotcycle-dat…
English
1
2
1
73
Efimia Panagiotaki retweetledi
Daniele De Martini
Daniele De Martini@d_d_martini·
Proud of our team: 5 papers accepted to #IROS2025 in Hangzhou! 🤖🇨🇳 We’re especially excited that one is a candidate for the IEEE-RAS Outstanding Women in Robotics & Automation (WiRA) Paper Award. Let’s meet up and talk robots! 👇
English
1
2
1
127
Efimia Panagiotaki
Efimia Panagiotaki@panefimia·
Having watched ~ an infinite number of @PetarV_93 ‘s tutorials, I can confidently say that this should not be missed. It opens a new door to the way we approach learning. For 🤖: NAR combines algorithms with DL so we don’t have to choose. It just works and it works every time🚀
Petar Veličković@PetarV_93

In < 1h, we will be taking the LoG stage for the new and improved version of our NAR tutorial 🔢! Hope you can join us -- it is open for all, publicly streamed on YouTube, and will feature a fun discussion of great reasoning research over both graphs 🕸️ and language 💬!

English
0
0
4
275
Efimia Panagiotaki retweetledi
Jacob Bamberger
Jacob Bamberger@jacobbamb·
We are happy to announce the Oxford LoG meet-up, LoG-Ox! The event is free to attend and will happen on November 25th. We plan to have keynote talks, lightning talks, posters and socials. More information and signup form is available here: log-ox.github.io. @LogConference
Jacob Bamberger tweet media
English
2
16
60
7.5K
Efimia Panagiotaki
Efimia Panagiotaki@panefimia·
Our method introduces a more interpretable and efficient robot learning framework. By integrating classical algorithms with deep learning, we combine structured reasoning and logical computations with the adaptability and generalisation capabilities of neural networks. 🚀 4/4
English
0
0
5
216
Efimia Panagiotaki
Efimia Panagiotaki@panefimia·
ICP algorithms: ❎require strong initialisation ❎sensitive to large translation errors ❎limited to Cartesian inputs ❎do not generalise NAR-*ICP: ✅superior performance ✅robust to noise/large errors ✅handle abstract input features ✅2-16x faster ✅differentiable 3/4
English
0
0
4
180
Efimia Panagiotaki
Efimia Panagiotaki@panefimia·
We extend the Neural Algorithmic Reasoning framework to address a fundamental robotics task. Our results showed that NAR-*ICP consistently outperforms even the algorithms it was trained on, especially in complex, noisy datasets where the traditional methods fail to converge. 2/4
English
0
0
3
200
Efimia Panagiotaki
Efimia Panagiotaki@panefimia·
We leverage the intermediate computations of ICP-based algorithms as supervisory signals to iteratively train GNNs to approximate the algorithmic reasoning processes, effectively combining neural networks and classical robotics algorithms for enhanced robot learning. 1/4
English
0
0
5
316
Efimia Panagiotaki
Efimia Panagiotaki@panefimia·
Attending the “Beyond the Symbols vs Signals Debate” scientific discussion meeting at the @royalsociety in London. Loads of great talks and insights in Neurosymbolic AI, Neuroscience, Reinforcement Learning, Representations, and Alignment. Also streamed online
Efimia Panagiotaki tweet media
English
1
0
3
151
Katerina Margatina
Katerina Margatina@katemargatina·
🦋Life updates! 1/ I finally got my PhD (with no corrections🥹)!!! Can’t thank enough my advisor @nikaletras for his support in the last 4.5 years. 🙏🏻 2/ Since Jan. 2024 I’ve joined the AWS Bedrock Agents team in NYC as an applied scientist!
Nikos Aletras@nikaletras

Congratulations to Dr. Katerina (⁦@katemargatina⁩) who passed her PhD viva today with no corrections!! 🎉🥳 Katerina has made very important contributions to active learning for NLP, including contrastive-based acquisition!

English
18
1
116
13.2K

Keşfet

@PetarV_93 @LogConference @d_d_martini @kula78 @royalsociety @katemargatina @nikaletras @MRG_Oxford