We created a new class of AI to accelerate research + discovery

Explicability as the foundation of ethical AI
Everyone seems to be talking about trustworthy AI these days. Yet it isn’t always clear what that is supposed to mean.

Abzu’s Summer 2022 Beer, Data, + Drugs event
A recap of Abzu’s Summer 2022 Beer, Data, + Drugs: An invite-only event to celebrate working in drug discovery

Peptide drug development with symbolic regression
An example of peptide drug featurization and modelling using anticancer peptides.

Evaluating the calibration of your QLattice models
Let’s study how well calibrated the QLattice models are, and to what extent calibrators can improve them.

An introduction to calibration (part I): Understanding the basics
What do machine learning model outputs represent? What if the predictions that we’re making come with a future risk?

An introduction to calibration (part II): Platt scaling, isotonic regression, and beta calibration
Calibrators are tools used to transform the scores generated by your models into (almost) real mathematical probabilities.

The European approach: Towards trustworthy AI
Something is brewing in Europe. Is it an ecosystem of excellence?

Techtopia podcast: Danish AI against breast cancer
Just because you can predict what’s going to happen does not mean you have an explanation for the phenomenon.

Understanding the data we create
We have to understand what the decisions we make are based on, not blindly trust that the computer is right.

The Healthtech Podcast: The Story of Abzu with Casper Wilstrup
Casper founded Abzu with seven brilliant, friendly, quirky humans. In this Healthtech Podcast he talks about everything healthcare and technology

Abzu’s keynote at TechBBQ 2021
Accelerating scientific discoveries with Abzu’s explainable AI: A breast cancer example

Multi-omics analysis made easy: A breast cancer example
The QLattice, a new explainable AI algorithm, can cut through the noise of omics data sets and point to the most relevant inputs and models.

siRNA molecule activity: explainable AI models uncover underlying mechanisms
Our results-driven world pressures us to give up good science.

Abzu’s artificial intelligence is fighting false news
The artificial intelligence deep tech startup identifies false news with its proprietary QLattice

How I Learned to Stop Encoding and Love the QLattice
Curing your data preprocessing blues: Automatic handling of categorical data and scaling

Meet Abzu, a European artificial intelligence company, where there are no bosses and everyone chooses their own salary
Abzu has chosen to take a unique approach to their business operations: There isn’t a formal hierarchy, every employee has a stake in the company,

Black-box medicine is taking over. Here’s an alternative.
The increasing application of black-box models gives rise to a range of both ethical and scientific questions.

AI and nucleic acid-based medicine: can symbolic regression unite them?
How the QLattice explains the cellular toxicity of RNA-targeting drugs.

The QLattice: A new machine learning model you didn’t know you needed
Data sciencey-sphere, I have big news. A radical new machine learning model has surfaced.

A new kind of AI
Do you think about machine learning? How about all the research put into self-driving cars or image recognition or natural language processing?

How Abzu adapted to Corona times
Small and large adjustments from a global artificial intelligence startup March 2020 is a month we will never forget. The world changed quite quickly, and
What people are saying








Publications + preprints
Combining symbolic regression with the Cox proportional hazards model improves prediction of heart failure deaths
Wilstrup, C., Cave, C. BMC Medical Informatics Decision Making 22, 196 (2022).
Symbolic regression analysis of interactions between first trimester maternal serum adipokines in pregnancies which develop pre-eclampsia
Casper Wilstrup, Paula L. Hedley, Line Rode, Sophie Placing, Karen R. Wøjdemann, Anne-Cathrine Shalmi, Karin Sundberg, Michael Christiansen
Explainable “White-box” machine learning is the way forward in pre-eclampsia screening
Michael Christiansen, MD, FRCPath, Casper Wilstrup, Paula L. Hedley, PhD, MPH. American Journal of Obstetrics and Gynecology, S0002-9378 (2022).
Identifying interactions in omics data for clinical biomarker discovery using symbolic regression
Niels Johan Christensen, Samuel Demharter, Meera Machado, Lykke Pedersen, Marco Salvatore, Valdemar Stentoft-Hansen, Miquel Triana Iglesias. Bioinformatics (2022).
Explainable long-term building energy consumption prediction using QLattice
Simon Wenninger, Can Kaymakci, Christian Wiethe. Applied Energy 308, 0306-2619 (2022).
Symbolic regression outperforms other models for small data sets
Casper Wilstrup, Jaan Kasak
HELLO, FOLKS
Contact Abzu
We’re just a bunch of nice nerds building something new and awesome. How can we help you?