Fink@World

8 juin 2026, 09:00 → 12 juin 2026, 18:30 Europe/Paris

Institut Pascal

Anais Möller (Swinburne University), Damien TURPIN (CEA-Saclay), Emille Ishida (CNRS/LPC-Clermont), Fabian Schussler (CEA/Irfu), Federica BRADASCIO (IJCLab, Université Paris-Saclay), Julien Peloton (CNRS-IJCLab), Marine HEBERT (LPC), Sarah ANTIER (IJCLAB), Vesna Cupic (IPa)

Welcome to Fink@World!

 

After 7 years of continuous learning and improvement with ZTF data, in February  2026 Fink has successfully started to stream alerts from the Vera C. Rubin Observatory. 

To celebrate this milestone, present first results, as well as to plan the next years of Fink's development, we hope to see you all in the south of Paris.

This is our first world-wide Fink event! 

The program will host a number of contributed talks and plenty of time for collaborative work whether you have an existing science analysis or a future development. We will also have discussions about our future plans.

 

In addition, every morning for 2 hours, we will run a school where you can learn about cloud computing, machine learning, and Fink/LSST data services. 

Note that talks will be streamed but remote presentations are not envisaged.

The meeting will happen at Institut Pascal - Orsay, France , between 8 - 12 June 2026.

There is no registration fee. 

Our venue holds a maximum of 60 participants, thus preference will be given for students, those currently working in Fink and active users. 

Asteroid_poster.pdf

FinkWorld_poster_for_printing.pdf

FinkWorld_without_marks.pdf

lun. 8 juin

Community: On-boarding

1 Registration

Logistics

2 Welcome

Orateur: Giuseppe Foffi (IPa)

School/Tutorial:: Introduction to Fink

Président de session: Dr Emille Ishida (CNRS/LPC-Clermont)

3 Fink until now

Orateur: Dr Julien Peloton (CNRS-IJCLab)

Introduction

4 Round table

Break: Coffee break

School/Tutorial:: The Portal

Président de session: Dr Emille Ishida (CNRS/LPC-Clermont)

5 The Fink Portal

Orateur: Dr Julien Peloton (CNRS-IJCLab)

Fink/LSST Portal

Break: Lunch

Talks:: Contributed talks

Président de session: Dr Emille Ishida (CNRS/LPC-Clermont)

6 Early science with Rubin and Fink

Orateur: Dr Anais Möller (Swinburne University)

AMoller_Finkworld.pdf

7 LSST Alert pipeline

Orateur: Bruno Sanchez (CPPM - CNRS)

LSST Alert Pipeline

15:10 Mini-break

8 Photometric classification of core collapse supernovae to determine luminosity functions at high redshifts.

With millions of detections per night the Vera C. Rubin Observatory (Rubin) will detect millions of supernovae (SNe) over the next ten years. This dataset presents a great opportunity to characterize core-collapse supernovae (SNCC) at redshifts > 0.2. Especially rates and properties of SNCC at these distances have been challenging to measure statistically until now.

I present a machine learning approach for the photometric classification of SNCC into their broad subtypes (SNII, SNIb/c). In preparation for Rubin, I use the Dark Energy Survey (DES) as a benchmark. The classifier is trained on simulations and tested on DES observations. My classifier achieves an accuracy of over 90% on the simulations with contamination below 5%. Validation of the method using low redshift SNCC samples from DES yields contamination below 20%. In this talk, I will present SNCC population properties at redshifts spanning 0.1 < z < 0.7 from DES, effectively extending the SNCC luminosity function and rate measurements to higher redshifts with a more diverse sample compared to the literature. I will conclude by discussing the implications of photometric SNCC classification and property estimations in the era of Rubin.

Orateur: M. Lukas Steinwender (Swinburne University of Technology)

CC_SN

temp_SteinwenderPhotClfSncc-1.pdf

9 Real-time photometric classification of superluminous supernovae

Superluminous Supernovae (SLSNe) are rare and extremely bright stellar explosions. Their precise powering mechanism remains an open question, with models suggesting contributions from magnetars, pair-instability explosions, or interaction with dense circumstellar material. Modern wide-field optical surveys, such as the Zwicky Transient Facility (ZTF), have allowed the automatic photometric detection of many SLSNe events, contributing greatly to the constraints on our models. The Vera C. Rubin Observatory (LSST) will significantly increase the number of SLSNe candidates by probing our Universe more deeply. It is expected to detect 104 SLSNe every year, which will deeply improve our knowledge on them.

In such big data surveys, efficiently identifying SLSNe among the vast number of transient alerts is very challenging and tailored pipelines should be constructed. Ideally, they should be fast enough to process large amounts of data, and interpretable enough to be trusted and used by the community.
We address this challenge by using a curated sample of spectroscopically confirmed SLSNe to develop a machine-learning classification framework for ZTF data. The model has been integrated into the Fink ecosystem, where it has already been operating in real time on ZTF alerts and has shared candidates publicly in dedicated channels. Over the past months, the classifier has demonstrated very high completeness and has assisted domain experts in identifying SLSN events. In this contribution, we present the methodology, performance, and limitations of the model, and we also discuss future prospects. In particular, its integration within the Fink environment facilitates quick and straightforward adaptation for the upcoming LSST era, which will represent a crucial period for scientific discoveries.

Orateur: Dr Etienne Russeil (Stockholm University)

NOMAI Fink@World.pdf

10 Updates on CATS

Orateur: Dr Bernardo Fraga (CBPF)

CATS

Break: Coffee break

Discussions: Plans for SN science with LSST alerts

Présidents de session: Dr Anais Möller (Swinburne University), Etienne Russeil (Stockholm University)

mar. 9 juin

Community: Change of venue

11 Meet at Institut Pascal

12 walk to Hbar

625 Rue Louis de Broglie, 91400 Orsay

Google map link

School/Tutorial:: The API

Président de session: Dr Julien Peloton (CNRS-IJCLab)

13 The API and SN

Orateur: Dr Anais Möller (Swinburne University)

Break: Coffee break

Talks:: Contributed talks

Président de session: Dr Julien Peloton (CNRS-IJCLab)

14 Optical-ɣ-Ray Correlations in Blazars from Time-Domain Surveys

Blazars are among the most variable non-thermal sources in the Universe, exhibiting broadband emissionfrom radio to γ-rays. With the new era of large-scale surveys such as the Vera C. Rubin Observatory, their optical variability can now be regularly monitored with unprecedented depth, paving the way to probe their emission mechanisms in a time-domain, multi-wavelength context.

In this work, we investigate the connection between optical and γ-ray emissions in a large sample of blazars using multi-year light curves from Fermi-LAT and Rubin’s predecessor, ZTF. We introduce a robust similarity metric to quantify cross-band correlations and implement a flare detection pipeline based on the rise and fall structure of extreme-amplitude emission events. Building on this, we develop a real-time algorithm designed to identify extreme optical states and trigger follow-up observations.

We find a zero time lag between the optical and γ-ray band for most of the sources, and optical-γ-ray correlations beyond > 3σ for about 20% of them, which supports co-spatial emission regions.

Our real-time triggering strategy achieves a purity of over 70% for γ-ray flares and nearly 100% for optical low states, demonstrating that optical surveys can efficiently anticipate high-energy activity and confidently trigger spectroscopic observations of the host galaxy. These results highlight the growing potential of allsky optical surveys as drivers of multi-wavelength follow-up, providing a powerful complement to current high-energy facilities such as Fermi-LAT.

Orateur: Julian Hamo (IJCLab)

2026.06.09_Fink@World.pdf

15 Hostless science module

Orateur: M. Rupesh Durgesh (COIN)

ELEPHANT

12:10 Mini-break

16 Early TDE science module

Orateur: Sergey Karpov (Institute of Physics, Czech Academy of Sciences)

early_tdes_lsst.pdf

17 A new microlensing classifier for Fink

Microlensing is a type of gravitational lensing phenomenom that occurs in scales of stars and planets. It manifests as an achromatic increase in the brightness of a star due to a stellar companion or even an orbiting planet, being one of the methods to find exoplanets.
Although existing simulated datasets are an incredible asset for training machine learning models, they lack diversity for some transients. We aim here to build a comprehensive simulated microlensing lightcurve dataset, including several different types of microlensing and contaminants, expanding on PLAsTiCC and ELAsTiCC. We train a classifier to identify microlensing events among the contaminants, but also one that is able to identify which type of system produced the signal. These models will be implemented in Fink in the near future

Orateur: Bernardo Fraga (CBPF)

microlensing

Break: Lunch

Talks:: Contributed talks

Président de session: Dr Anais Möller (Swinburne University)

18 Icare/Skyportal

Orateur: Camille DOUZET (IJCLab)

Fink@World Skyportal _ ICARE-4.pdf

19 BHTOM: more than a Target Observation Manager

BHTOM (Black Hole Target Observation Manager) is a web-based platform for coordinating time-domain astronomical observations across a global network of telescopes. It enables users to define targets, collect and process photometric data automatically, and combine new observations with archival datasets. The system standardizes data using PSF photometry and Gaia-based filters. Designed for researchers, telescope operators, educators, and amateurs, BHTOM supports collaborative, scalable, and data-driven astronomy, particularly suited for the era of large survey facilities like the Vera C. Rubin Observatory.

In this talk, I will present the key features of BHTOM and demonstrate how it enables LSST synergies by integrating follow up with archival data (the platform combines LSST and ZTF observations in a single interface for example), maximizing scientific return for time-domain studies.

Orateur: Dr Priscila Pessi (National Center for Nuclear Research (NCBJ), Warsaw)

Pessi_Fink@World2026.pdf

15:10 Mini-break

20 Astro-COLIBRI

The Astro-COLIBRI platform is a comprehensive ecosystem designed for the real-time evaluation and coordination of transient astrophysical events. It integrates diverse alert streams, including neutrinos, gravitational waves, and high-energy gamma rays, into a unified, user-friendly interface.
The link between Astro-COLIBRI and Fink is a critical bridge in the multi-messenger pipeline: Fink processes the "big data" stream from LSST to identify promising candidates (such as kilonovae, TDEs or early-phase supernovae), which are then ingested by Astro-COLIBRI via the Transient Name Server and (in the future) real-time Kafka streams and APIs. Astro-COLIBRI contextualizes these alerts with multi-wavelength archival data and provides observability assessments, enabling the global community to trigger rapid follow-up observations within minutes of a detection.

Orateur: Fabian Schussler (CEA/Irfu)

Astro_COLIBRI_2026-06-09.pdf

21 KNC @Fink: Connecting Fink/LSST transients to the amateur world

Since 2018, led by the GRANDMA Collaboration, the Kilonova-Catcher (KNC) citizen science program aims to directly involve a worldwide community of citizens in the search for electromagnetic counterparts of gravitational-wave sources. Over the past two years, the scientific scope of KNC has broadened to include “fast” extragalactic transients such as gamma-ray burst afterglows, nearby supernovae, cosmological GRB-SNe, cataclysmic variables, and more.

The KNC community is growing and can now support several observational programs simultaneously, significantly contributing to time-domain astronomy discoveries and follow-up campaigns. In this presentation, I will provide an overview of the current status of the KNC network and its organizational structure. I will also present insights into the performance of this worldwide network of citizen astronomers and relate them to the needs of the transient astronomy community using Fink data.

Orateur: Damien TURPIN (CEA-Saclay)

KNC_FINK@world.pdf

Community: Group picture

Break: Coffee break

Discussions: Connections at scale: LSST and curent services

Présidents de session: Dr Julien Peloton (CNRS-IJCLab), Dr Priscila Pessi (National Center for Nuclear Research (NCBJ), Warsaw)

Public talk

Président de session: Dr Julien Peloton (CNRS-IJCLab)

22 Impacts d'astéroïdes: risque réel ou hypothétique?

Les asteroïdes sont les restes des premiers solides qui se sont agglomérés pour former les planètes au début du système solaire. Des millions de ces corps rocheux, allant de quelques mètres de diamètre à plusieurs centaines de kilomètres, se trouvent encore en orbite autour du Soleil, principalement entre Mars et Jupiter.

Une partie d'entre eux se situe sur des orbites croisant celle de la Terre. Occasionnellement, ils la percutent, déposant chaque année plusieurs milliers de tonnes de météorites à la surface du globe. Si ces chutes sont généralement sans danger, l'impact d'un astéroïde kilométrique serait une catastrophe (les dinosaures ne peuvent plus en témoigner).

Nous verrons comment les astronomes scrutent le ciel afin de découvrir ces corps, et comment leur trajectoire est étudiée pour identifier les astéroides potentiellement dangeureux. Finalement, nous verrons les solutions à l'étude si un impact devait se produire

Orateur: Dr Benoit Carry (Lagrange, Observatoire de la Cote d Azur)

mer. 10 juin

School/Tutorial:: Solar System tools in Fink

Président de session: Dr Julien Peloton (CNRS-IJCLab)

23 Solar System tools

Orateur: Roman Le Montagner

Break: Coffee break

Talks:: Contributed talks

Président de session: Dr Anais Möller (Swinburne University)

24 Shape, Orientation and Colors Combined Algorithm (SOCCA)

Large photometric surveys provide sparse multi-band lightcurves for millions of Solar System objects (SSOs),
offering an opportunity to jointly constrain their physical and compositional properties. However, current phase func-
tion models do not account for rotational variability, limiting their ability to retrieve accurate parameters. Similarly,
methods that recover shape and rotational parameters remain computationally expensive, making the extraction of
such properties prohibitive at scale.
We aim to develop a model capable of simultaneously retrieving the absolute magnitude, phase parameters, spin
state, and shape proportions of SSOs from sparse photometric data, while remaining efficient for large datasets.
We introduce the Shape, Orientation and Colors Combined Algorithm (SOCCA), which extends the HG1 G2
formalism by incorporating the projected surface of a rotating triaxial ellipsoid. The model jointly fits multi-band pho-
tometry, and includes a dedicated treatment of rotational period determination. We validate the method on simulated
LSST-like observations and on a real lightcurve.
SOCCA significantly improves the fit to photometric data, reducing the mean residuals to half, compared to
previous models. It retrieves the absolute magnitude with a scatter about three times smaller than existing approaches,
and improves the determination of phase parameters by a similar factor. The inclusion of shape and rotation increases
the number of valid solutions by ∼10-20% per filter, leading to an overall success rate of 50%.
By combining phase, shape, and rotational information in a single model, SOCCA provides a more complete
physical description of SSOs from sparse photometry. Its performance and scalability make it well suited for current
and upcoming large surveys such as the Zwicky Transient Facility (ZTF) and the recently launched Legacy Survey of
Space and Time (LSST).

Orateur: Konstantinos Odysseas Xenos (Observatoire de la Cote d'Azur)

main.pdf

25 Photometric Follow-Up of LSST Ultra-Fast Rotators in the Main Asteroid Belt

The LSST is revealing a previously unexplored population of ultra-fast-rotating sub-kilometer asteroids in the main belt. Several recently identified targets, including 2025 MN45, 2025 MK41, 2025 MV71, and 2025 MG56, show rotation periods of just a few minutes despite sizes >0.5 km, challenging current rubble-pile cohesion models. We aim to perform targeted photometric follow-up to verify rotation periods, characterize lightcurves, and improve size estimates by combining observations at opposition with LSST data. This work will clarify their physical properties, internal structures, and formation histories, providing critical insights into asteroid cohesion limits and spin evolution.

Orateur: Milagros Colazo (Adam Mickiewicz University)

Colazo_FINK@World26.pdf

12:10 Mini-break

26 Detecting Active Solar System Objects in the Fink Alert Stream

Predictions suggest that the Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST) will discover ~six million Solar System objects (SSOs) in its decade of operations. The onset of activity in an SSO, such as coma or tail formation, signals a change in its physical environment that can also influence its orbit. Rapid identification of such activity is therefore essential both for understanding the physical processes driving SSO evolution and for improving orbital solutions through timely follow-up observations.

However, there are currently no dedicated tools designed to identify active SSOs within the LSST alert stream. I present ongoing work to develop a machine-learning pipeline that could potentially be part of the Fink broker ecosystem to flag candidate active SSOs. The approach combines anomaly detection with neural-network classifiers to identify alerts whose photometric or morphological features deviate from those of typical inactive asteroids. An active learning strategy is used to iteratively refine the model by prioritising uncertain or anomalous alerts for human labelling, enabling efficient training in a regime where labeled examples of active objects remain scarce.

This work aims to provide an early-warning capability for identifying active SSOs within the Fink alert stream, enabling rapid follow-up observations and improving our ability to study and track these dynamically evolving objects.

Orateur: Dr Preeti Cowan (University of Auckland)

FinkWorld26_PCowan.pptx

27 Towards Real-Time Asteroid Tracking with Fink-FAT and Outfit

Fink-FAT (Fink Asteroid Tracker) is a Rust-based pipeline for ingesting photometric alerts, linking detections of moving objects across one or more nights, and reconstructing candidate trajectories. Outfit, a standalone Rust package, is used within this workflow to derive preliminary orbital solutions. In this talk, we will first introduce Outfit and Fink-FAT, then present the first results obtained from early alert data, with a focus on algorithmic performance, robustness, and current limitations. We will conclude with a live demonstration of the pipeline, illustrating how the two tools can support the analysis of high-cadence astronomical alerts.

Orateur: Roman Le Montagner

fink_fat_asteroid_linking.pdf

Break: Lunch

Discussions: Solar System Science with Rubin: potential and scientific synergies

Présidents de session: Dr Benoit Carry (Lagrange, Observatoire de la Cote d Azur), Preeti Cowan (University of Auckland)

Break: Coffee break

Community: Brain Storming Fink Dream Shots

Président de session: Dr Emille Ishida (CNRS/LPC-Clermont)

Break: Happy hour

jeu. 11 juin

School/Tutorial:: Live Stream and Fink Bots

Président de session: Dr Emille Ishida (CNRS/LPC-Clermont)

28 Livestream & bots

Orateur: Camille DOUZET (IJCLab)

Fink Filters .pdf

29 Classification graphs to analyze alerts

Orateur: Julius HRIVNAC

Graphs

Break: Coffee break

Talks:: Contributed talks

Président de session: Dr Anais Möller (Swinburne University)

30 Automated Physical Inference on Fink Alerts Using Neural Networks

Inverse problems in astronomy are often computationally expensive, and Markov Chain Monte Carlo (MCMC) methods become impractical when dealing with massive, heterogeneous datasets or when only limited observations are available. With the advent of the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST), the astronomical community is expected to receive millions of transient alerts per night, many of which evolve on very short timescales. Rapid classification and characterization of these events is therefore essential.

In this work, we present a Physics-Informed Neural Network (PINN) framework to model the historical light curves of Fink alerts. Using the Type II-P supernova SN~2022acko as a test case, we demonstrate that our method can accurately infer the physical properties of the transient progenitor system, achieving reliable results within only a few days of observations. In contrast to traditional MCMC approaches, this method requires minimal human intervention, enabling scalable, automated, and physics-driven characterization of large volumes of transients in the Fink alert stream.

Orateur: Dr Gabriel Teixeira (CBPF)

PILCA

31 LStein: A new approach to visualizing Rubin Lightcurves

Extracting the full scientific potential of large photometric surveys demands richer data visualization than current methods provide. The Vera C. Rubin Observatory's LSST exemplifies this challenge: its filter-based sampling of the light spectrum is intentionally sparse, yet the subtle details encoded in filter spacing, overlap, and sensitivity profiles carry critical information that standard lightcurve visualisation discards. Recovering this information unlocks new opportunities across high-impact science cases such as supernova classification, asteroseismology, and exoplanet research.

I present LStein (Linking Series to envision information neatly), a novel, open-source, deterministic visualisation. LStein addresses this visualization challenge by including the missing information through projection: A set of lightcurves is placed in a combined angular coordinate system using azimuthal offsets to encode the sparse wavelength dimension. Using LStein one can gain intuitive insight into the relation between passbands while still showing all individual lightcurves in a comparable manner. This enables the researcher to include the wavelength information in interpretations of the physical nature of signals. I will demonstrate the capabilities of LStein by applying it to the first Rubin alerts. Examples include supernovae, and variable stars from Fink broker. I will conclude by outlining extensions to other fields such as radio astronomy and spectral analysis.

Orateur: Lukas Steinwender (Swinburne University of Technology)

temp_SteiwenderLstein.pdf

12:10 Mini-break

32 Origin of Dipole Features in Rubin Alerts: A Residual effect from differential atmospheric Refraction

We investigate the origin of dipole features observed in some Rubin alert light curves, as identified in the Fink alert broker stream. These dipoles, notably present in cross-matched sources such as Gaia counterparts, are characterized by a dipole angle and an angular separation provided at the alert level.
Our analysis reveals a strong angular dependence of these features, suggestive of a systematic effect rather than an astrophysical signal. In particular, the observed behavior is consistent with expectations from parallactic effects induced by atmospheric refraction, which can introduce chromatic centroid shifts depending on observing geometry.
We test this hypothesis by comparing the measured dipole orientations and amplitudes to the expected parallactic angle and its associated scaling with observational parameters. We present the current status of this study, including quantitative comparisons and limitations of the available alert-level information. Finally, we discuss the implications for alert stream analysis and broker-level filtering, as well as perspectives for improving the modeling and mitigation of such effects in Rubin data products.

Orateur: Dr Sylvie Dagoret (IJCLab)

2026-06-11-FinkAtWorld_dagoret_final.pdf

33 Learning from Rubin first alerts

The release of Rubin first alerts is a great opportunity to get acquainted with the data. Using anomaly finder as a driver, I'll showcase the first lesson learnt. This will concern both data (what are they, are there important flags linked to data quality ? What to do with forced photometry...) and maybe if lucky first anomalies.
Obviously, this is work in progress for now, I can't promise yet what I'll find as it will depend on pending work. However, I hope this will be a great opportunity to trigger discussion with other Rubin data enthousiasts...

Orateur: Emmanuel Gangler (LPC)

PresentationJune26.pdf

34 Anomaly paper

Orateur: Dr Emille Ishida (CNRS/LPC-Clermont)

Anomaly paper

Break: Lunch

Break: Free time

Break: Sightseeing: Musee d'Orsay

Break: Conference dinner

ven. 12 juin

School/Tutorial:: Science module integration

Président de session: Dr Emille Ishida (CNRS/LPC-Clermont)

35 Science modules

Orateurs: Farid MAMAN (IJCLAB), Massinissa MACHTER (CNRS/IJCLAB/ing-dev-informatique)

Fink_AI_Presentation.pptx-5.pdf

Fink@world - profiling science modules.pdf

Break: Coffee break

Discussions: The future of Fink

Présidents de session: Dr Anais Möller (Swinburne University), Dr Emille Ishida (CNRS/LPC-Clermont), Dr Julien Peloton (CNRS-IJCLab)

36 Fink in the next decade

12:10 Mini-break

37 Dream Shot discussion

Break: Lunch

Community: Office hours