Sketching the Future

Workshop on Advanced Data Sketching Tools

About

Modern data-intensive applications continuously generate massive streams of data, from network monitoring and cloud infrastructure to digital twins and AI systems. Analyzing this data in real time is often impossible using exact methods alone due to memory and latency constraints. Synopses, including sketches, are compact data summaries that enable fast, memory-efficient approximate analytics with provable accuracy guarantees. They are already widely deployed in modern data systems, yet many powerful synopsis techniques remain underutilized, while emerging AI applications present new opportunities and challenges. This interactive workshop is driven by real-world use cases, bringing together researchers and practitioners to identify where current methods fall short and where new synopsis techniques can have the greatest impact.

The workshop brings together researchers in algorithms, databases, distributed systems, and AI/ML with industry practitioners building large-scale data platforms and AI systems. Through interactive discussions centred on practical use cases, participants will identify high-impact research challenges, discuss barriers to industrial adoption, and shape a shared research agenda for the next generation of synopsis techniques.

Our goals are to:

  • Bridge the gap between research and practice.
  • Identify open problems that will shape the field over the coming decade.
  • Develop a shared vision for the future of the data sketching community through interactive discussions and collaborative sessions.
  • Foster lasting collaborations between researchers and practitioners.

The outcomes of the workshop will be summarized in a post-workshop white paper.

Practical Information

Location

Lorentz Center, Leiden
Niels Bohrweg 1, 2nd floor, 2333 CA Leiden, The Netherlands

Date

20 January 9:00 to 22 January 17:30 2027

Costs

The workshop is free of charge and invitation-only. We expect everyone to cover their own travel and accomodation costs.
The Lorentz Center offers accomodation for participants at a reduced rate. The costs of this hotel are expected to be around 137 euros for one night, including breakfast and taxes.
More information will follow.

Travel

Leiden has a train connection with Schiphol Airport of approximately 20 minutes, and with Amsterdam Central of 37 minutes.

Adjacent events

The workshop is hosted in the week preceding the Conference on Innovative Data systems Research (CIDR) that is taking place in Amsterdam.

Organizers

The workshop is organized by:
Wieger Punter (TU Eindhoven). E-mail: w.r.punter [at] tue.nl. Link to webpage.
Odysseas Papapetrou (TU Eindhoven). E-mail: o.papapetrou [at] tue.nl. Link to webpage.
Minos Garofalakis (Athena RC, TU Crete). E-mail: minos [at] athenarc.gr, Link to webpage.

Questions?

For questions, don't hesitate to contact w.r.punter [at] tue.nl.

Program

Below is the full tentative program for the 3-day workshop.
Before the onsite workshop, there is an online preparation meeting where we will discuss expectations of the workshop, briefly present the use cases, so that the time during the onsite workshop is used as efficient as possible.

Day 1 - 20 Jan

Coffee and registration 9:30 - 10:00
Introduction & Welcome Lorentz Center 10:00 - 10:15
Introduction Goals Organizers 10:15 - 10:30

Overview of the workshop goals, structure, and expected outcomes.

Keynote: Sketches Over the Decades 10:30 - 11:15
Participant Introductions 11:15 - 11:30
Panel Discussion: Community Progress 11:30 - 12:00

Discussion on what the community has achieved, which research reached production, what inspired follow-up work, and why some ideas did not make it. This is to set the flavor of the remaining work- shop.

Lunch Break 12:00 - 13:30
Industry Use Cases 13:30 - 15:00

Presentations and Q&A on industry use cases: (1) Data-intensive science, (2) control systems monitoring and (3) sketching techniques for AI

Coffee Break 15:00 - 15:30

Formation on breakout groups on use cases.

Breakout Sessions: Use Cases 15:30 - 17:00

Group discussions focusing on specific use cases and identifying open problems.

Reception 17:00 - 19:00

Informal networking and continued discussions.

Day 2 - 21 Jan

Opening & Recap Day 1 9:00 - 9:30
Breakout Sessions: Solution exploration 9:30 - 10:30

Groups discuss potential solutions to identified problems.

Coffee Break 10:30 - 11:00
Plenary Discussion on solutions 11:00 - 11:30

Identify abstractions of the problems.

Breakout Sessions: solutions and abstractions 11:30 - 12:00

Groups discuss solutions to identified problems and abstractions to other problem settings.

Lunch Break 12:00 - 13:30
Breakout Sessions: solutions and abstractions 13:30 - 14:00

Groups discuss solutions to identified problems and abstractions to other problem settings.

Midday Recap 14:00 - 14:30

Groups share interim findings and insights.

Breakout Presentations and Q&A 14:30 - 15:45

Groups present results to each other.

Coffee Break 15:45 - 16:15
Plenary Discussion & Recap 16:15 - 17:15

Joint discussion and synthesis of the day's outcomes.

Dinner
18:00

Day 3 - 22 Jan

Keynote 9:00 - 9:45

Forward-looking keynote setting the stage for future research directions.

Q&A and discussion 9:45 - 10:150

New thoughts, reflections and open questions.

Coffee Break 10:15 - 10:45
Lessons from Use Cases 10:45 - 11:15

What have researchers and practitioners learned from discussing the use cases?

Vision statement discussion 11:15 - 12:00

Developing a shared vision for the field.

Lunch Break 12:00 - 13:30
Dissemination Planning 13:30 - 14:00

How to communicate the vision and results to the wider community?

Open Research Discussion 14:00 - 15:45

Focused discussion on concrete research problems and open topics.

Closing Statement 15:45 - 16:00

Summary of outcomes and next steps.

Drinks 16:00 - 17:30

Informal wrap-up of the workshop.

A month after the workshop, an online meeting will take place to recap the onsite workshop. The white paper on the vision for the community will be presented here, and open challenges will be discussed. Also, this session gives space to discuss topics that came to participants minds following the workshop.

Use Cases

During the workshop, we will discuss several use cases that are relevant for the community. The use cases will be presented by industry partners, and will be discussed in breakout sessions. The goal is to identify open problems and challenges that can be addressed by the community.

  • Data-intensive science
  • Control systems monitoring
  • Sketching techniques for AI

The partners that contribute these use cases, and a brief description, will be announced soon, as well as a list of confirmed participants.