Dominik Grimm

Happy to share that our paper, "Self-Improvement for Neural Combinatorial Optimization" (openreview.net/forum?id=agT8o…), published in #TMLR and awarded a 'Featured Certification', has been invited to be presented at #ICLR2025 in Singapore! @TmlrOrg @iclr_conf @bit_tumcs #TUMCS

GraphXForm: Graph transformer for computer-aided molecular design with application to extraction • GraphXForm introduces a novel graph transformer model for molecular design, focusing on applications in solvent design for liquid-liquid extraction processes. • This model tackles limitations of string-based molecular representations (e.g., SMILES) by directly modifying molecular graphs, which ensures chemical validity and allows structural constraints like specific substructures or bond restrictions. • Unlike traditional deep learning approaches, GraphXForm iteratively builds molecular graphs, predicting the next atom or bond action with a transformer trained on existing molecular structures. • Fine-tuning employs a unique training algorithm that combines deep cross-entropy and self-improvement learning, stabilizing the training of multi-layered graph transformers without reinforcement learning. • Tested on two solvent design tasks, GraphXForm outperforms four state-of-the-art models, including REINVENT and Junction Tree VAE, demonstrating superior objective scores in both efficiency and flexibility. • By integrating constraints such as limiting ring sizes or bond types, GraphXForm ensures designed solvents meet industrial specifications, highlighting its adaptability for complex molecular tasks. @dg_grimm 💻Code: github.com/grimmlab/graph… 📜Paper: arxiv.org/abs/2411.01667 #MolecularDesign #GraphTransformers #MachineLearning #ChemicalEngineering #SolventDesign

permGWAS2 has been published! It does not break the population structure during permutations and uses an elegant block matrix decomposition to optimize computations. It also yields lower FDR for skewed phenotypic distributions academic.oup.com/bioinformatics… @bit_tumcs #HSWT #TUMCS

🚀 We had an incredible time hosting the ML4CCE Workshop at @ECMLPKDD 2024! 💡🔬 Experts from machine learning and chemical engineering came together to explore exciting topics like molecular design, process control, and anomaly detection, among others. 🎓✨

Special thanks to our Keynote speakers: Venkat Venkatasubramanian (@ProfVenkat7), Dominik Grimm (@dg_grimm), and Felix Strieth-Kalthoff (@felix_s_k) for their incredible talks and insights! 🎤👏 Their contributions made the event truly memorable.

Our latest paper "Self-Improvement for Neural Combinatorial Optimization: Sample Without Replacement, but Improvement" published @TmlrOrg. Impressive work by my PhD student Jonathan Pirnay. Awarded with a "Featured" certification. openreview.net/forum?id=agT8o… @TU_Muenchen #TUMCS #HSWT

Given a invited talk today @ECMLPKDD at the Machine Learning for Chemistry and Chemical Engineering (ML4CCE) Workshop about „Automated flowsheet synthesis with deep reinforcement learning“ #ECMLPKDD24 #Vilnius #TUMCS #HSWT

In our latest @naturemethods perspective we provide 7 questions on how to detect data leakage in biological #MachineLearning applications. @judith_bernett @dbblumenthal @bit_tumcs @FloHasee @RomanJoeres @itisalist @TU_Muenchen #TUMCS #HSWT rdcu.be/dQuw1

Very proud to present our latest work: 7 guiding questions to avoid data leakage in biological machine learning applications ✨🔍 We hope that reflecting on these questions helps researchers to identify issues or shortcuts leading to overly optimistic performance estimates. 📈🧑‍🔬

The entire collection of Focus of advanced AI in biology pieces can also be found here: nature.com/collections/ah… We thank all of our authors for their fantastic contributions!

A Perspective from @itisalist @judith_bernett @RomanJoeres @ok55991 @FloHasee @dg_grimm @bit_tumcs & @dbblumenthal discusses the issue of data leakage in machine learning models and presents 7 questions to identify and avoid problems as a result. nature.com/articles/s4159…

It also allows the training of deep transformer models where reinforcement learning is too slow. While our paper presents our method from a general machine learning perspective, we are currently applying this approach to problems in chemistry, chemical engineering, and genomics!

In this work, we bridge these approaches with a novel strategy that trains the network on "pseudo-expert" output that it generates itself, without any external guidance. This is a modern approach that aligns with the recent success of self-improvement learning in reasoning LLMs.

Self-Improvement for Neural Combinatorial Optimization: Sample Without Replacement, but Improvement Jonathan Pirnay, Dominik G. Grimm. Action editor: Tim Genewein. openreview.net/forum?id=agT8o… #scheduling #imitation #salesman

Self-Improvement for Neural Combinatorial Optimization: Sample without Replacement, but Improvement. arxiv.org/abs/2403.15180

First paper of the year in Scientific Data. An impressive „Manually annotated and curated Dataset of diverse Weed Species in Maize and Sorghum for Computer Vision“ @NikitaGenze @bit_tumcs @TU_Muenchen #HSWT #TUMCS #PrecisionAgriculture nature.com/articles/s4159…

First publication out with @bit_tumcs!! Read this contribution about a #prediction model (#ProLaTherm) for #thermophilicsproteins. Protein thermostability is instrumental for protein #hybrid #optoelectronics! 👉academic.oup.com/nargab/article…