David Selby @davidselby.bsky.social
@TeaStats
Followers
405
Following
363
Media
140
Statuses
463
Enthusiastic about tea, statistics and t-statistics. Researcher in Data Science & its Applications @DFKI, honorary @CfE_UoM @PARADISE_AI. #Rstats evangelist
Kaiserslautern 🇩🇪
Joined June 2014
🧬BioDisco, an open-source biomedical hypothesis generator, uses agentic LLMs, knowledge graphs and literature search, with an iterative self-evaluation loop, significantly outperforming other architectures. Preprint:
arxiv.org
Identifying novel hypotheses is essential to scientific research, yet this process risks being overwhelmed by the sheer volume and complexity of available information. Existing automated methods...
0
1
1
New: unofficial @quarto_pub template for the upcoming @RealAAAI 2026 conference. Write your submission in Markdown with embedded computations!
0
0
0
RT @FrontComputSci: New Research: Visible neural networks for multi-omics integration: a critical review #Frontiers….
frontiersin.org
BackgroundBiomarker discovery and drug response prediction are central to personalized medicine, driving demand for predictive models that also offer biologi...
0
1
0
XAI, AutoML and perverse publishing incentives could create a perfect storm for "X-hacking": In this paper presented at @icmlconf, we describe a new threat to reproducible research and trustworthy AI:
dfki.de
At ICML 2025, DFKI researchers show how AutoML can generate misleading AI explanations - and propose new standards for trustworthy AI.
0
0
0
❓What is a "Visible Neural Network"? A new deep learning model for omics, where prior knowledge and interpretability are baked right into the architecture. 🎯 We review dozens of models, datasets & applications, and call for better tools/benchmarks:.
frontiersin.org
BackgroundBiomarker discovery and drug response prediction are central to personalized medicine, driving demand for predictive models that also offer biologi...
0
2
1
RT @cwcyau: Health Research From Home Hackathon 2025 |.This hackathon is being held by Health Research From Home Partnership led by the @Of….
health-research-from-home.github.io
7-9 May 2025
0
2
0
Lay abstract for our latest article on retrieving quantitative expert knowledge from LLMs
statisticsviews.com
The lay abstract featured today (for Had Enough of Experts? Quantitative Knowledge Retrieval From Large Language Models by David Selby, Yuichiro Iwashita, Kai Spriestersbach, Mohammad Saad, Dennis...
0
0
0
Just published! Can LLMs, having read so much scientific literature, play the role of a human expert and help us fill in missing values and fit statistical models to small data sets? We investigate:.
onlinelibrary.wiley.com
Large language models (LLMs) have been extensively studied for their ability to generate convincing natural language sequences; however, their utility for quantitative information retrieval is less...
0
1
2
New blog post: on learning new English words and meanings in Germany
selbydavid.com
At the railway station, a lost-looking US soldier asked me if I spoke English. Do I? At times it feels like it, but the Germans keep me guessing. Since moving to Germany, I have been continually...
0
0
0
New blog post: Alternatives to @overleaf for collaborative and reproducible writing by combining @code with @quarto_pub or #Rstats markdown.
selbydavid.com
Overleaf, formerly known as Share$\LaTeX$, is the go-to collaborative document editor for many researchers, who have taken advantage of its free tier. It’s a web-based editor that compiles $\LaTeX$...
0
0
0
Thrilled to share our latest publication in @NatureRevGenet!. We explore how deep learning models infused with prior pathway knowledge — aka 'visible neural networks' — promise better predictive accuracy & interpretability in multi-omics data analysis.
nature.com
Nature Reviews Genetics - Biologically informed neural networks promise to lead to more explainable, data-driven discoveries in genomics, drug development and precision medicine. Selby et al....
0
18
36
Friends don't let friends make dynamite plunger plots.
This might be the first time after 10 years that boosted trees are not the best default choice when working with data in tables. Instead a pre-trained neural network is, the new TabPFN, as we just published in Nature 🎉
1
0
1
Pleased to present our poster at #NeurIPS2024 workshop on Bayesian Decisionmaking and Uncertainty! 🎉 Our work explores using large language models for eliciting expert-informed Bayesian priors. Elicited lots of discussion with community too! Check it out:
0
1
5
New preprint on Visible Neural Networks, a way of integrating prior biological knowledge into machine learning models of omics data to improve interpretability. We review >80 papers to explain what VNNs are, how you build them and how they are evaluated.
biorxiv.org
Biomarker discovery and drug response prediction is central to personalized medicine, driving demand for predictive models that also offer biological insights. Biologically informed neural networks...
0
0
1
RT @joncstone: Every form of bike parking that isn’t just a normal Sheffield stand is worse than a normal Sheffield stand.
0
60
0
RT @WGDixon: First steps towards providing care at the right time for people living with #RA. Small exploratory analysis of predicting flar….
formative.jmir.org
Background: The ability to predict rheumatoid arthritis (RA) flares between clinic visits based on real-time, longitudinal patient-generated data could potentially allow for timely interventions to...
0
3
0
RT @data_sci_apps: We are looking for enthusiastic Data Science Researchers to join our growing team! Interested in trustworthy AI, spatio-….
jobs.dfki.de
0
3
0
RT @timleunig: Brilliant long read from @ft @Louis_Ashworth on the idiocy of VAT on flapjacks: "Two fully-grown men eating a year-out-of-da….
0
2
0