From 740eb17d2f647681e0f4906e38b010dab06d9b2c Mon Sep 17 00:00:00 2001 From: Kristian Kersting Date: Sun, 29 Dec 2024 15:50:35 +0100 Subject: [PATCH] update --- people/kkersting/index.html | 2 +- references.bib | 6 +++--- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/people/kkersting/index.html b/people/kkersting/index.html index f89af56..006bf82 100644 --- a/people/kkersting/index.html +++ b/people/kkersting/index.html @@ -226,7 +226,7 @@ co-founded the KI-Klub, connecting AI experts, media, public, and politicians, and published the Springer book Wie Maschinen Lernen, - one of the first German general introductory books on AI and, in particular, machine learning educating the promise and potential of AI to the broader society. We were also privileged enough to inform the public about AI and get our own work covered in (social) media such as New York Times, Washington Post, Financial Times, MIT Technology Review, Scientific American, ARTE, Frontiers Science Blog, Science, FAZ Digitec Podcast, KfW Podcast „Zukunft:digital“, Chainless Life Podcast, Heute Journal, Frankfurter Allgemeine Zeitung, Tagesspiegel, Handelsblatt, Frankfurter Rundschau, Zeit Campus, Heise, Spektrum der Wissenschaften, iX Magazin, Focus, among others, and an exhibition at the Nibelungen Museum in Worms, Germany, as well as a AI featured concert with the Singakademie Dresden e.V. To express my point of view on AI, I also had a regular column on AI in the German (Sunday) newspaper Welt (am Sonntag). + one of the first German general introductory books on AI and, in particular, machine learning educating the promise and potential of AI to the broader society. We were also privileged enough to inform the public about AI and get our own work covered in (social) media such as New York Times, Washington Post, Financial Times, MIT Technology Review, Scientific American, ARTE, Frontiers Science Blog, Science, FAZ Digitec Podcast, KfW Podcast „Zukunft:digital“, Chainless Life Podcast, Heute Journal, Frankfurter Allgemeine Zeitung, Tagesspiegel, Handelsblatt, Frankfurter Rundschau, Süddeutsche Zeitung, Zeit Campus, Heise, Spektrum der Wissenschaften, iX Magazin, Focus, among others, and an exhibition at the Nibelungen Museum in Worms, Germany, as well as a AI featured concert with the Singakademie Dresden e.V. To express my point of view on AI, I also had a regular column on AI in the German (Sunday) newspaper Welt (am Sonntag).

diff --git a/references.bib b/references.bib index 240a2e5..15ec5f0 100644 --- a/references.bib +++ b/references.bib @@ -11,9 +11,9 @@ @misc{friedrich2024llms @misc{skryagin2024asn, Anote = {./images/answer_set_networks.png}, - title={Answer Set Networks: Casting Answer Set Programming into Deep Learning}, + title={Answer Set Networks: Casting Answer Set Programming into Deep Learning}, author={Arseny Skryagin and Daniel Ochs and Phillip Deibert and Simon Kohaut and Devendra Singh Dhami and Kristian Kersting}, - Note={Although Answer Set Programming (ASP) allows constraining neural-symbolic (NeSy) systems, its employment is hindered by the prohibitive costs of computing stable models and the CPU-bound nature of state-of-the-art solvers. To this end, we propose Answer Set Networks (ASN), a NeSy solver. Based on Graph Neural Networks (GNN), ASNs are a scalable approach to ASP-based Deep Probabilistic Logic Programming (DPPL). Specifically, we show how to translate ASPs into ASNs and demonstrate how ASNs can efficiently solve the encoded problem by leveraging GPU's batching and parallelization capabilities. Our experimental evaluations demonstrate that ASNs outperform state-of-the-art CPU-bound NeSy systems on multiple tasks. Simultaneously, we make the following two contributions based on the strengths of ASNs. Namely, we are the first to show the finetuning of Large Language Models (LLM) with DPPLs, employing ASNs to guide the training with logic. Further, we show the "constitutional navigation" of drones, i.e., encoding public aviation laws in an ASN for routing Unmanned Aerial Vehicles in uncertain environments.}, + Note={Although Answer Set Programming (ASP) allows constraining neural-symbolic (NeSy) systems, its employment is hindered by the prohibitive costs of computing stable models and the CPU-bound nature of state-of-the-art solvers. To this end, we propose Answer Set Networks (ASN), a NeSy solver. Based on Graph Neural Networks (GNN), ASNs are a scalable approach to ASP-based Deep Probabilistic Logic Programming (DPPL). Specifically, we show how to translate ASPs into ASNs and demonstrate how ASNs can efficiently solve the encoded problem by leveraging GPU's batching and parallelization capabilities. Our experimental evaluations demonstrate that ASNs outperform state-of-the-art CPU-bound NeSy systems on multiple tasks. Simultaneously, we make the following two contributions based on the strengths of ASNs. Namely, we are the first to show the finetuning of Large Language Models (LLM) with DPPLs, employing ASNs to guide the training with logic. Further, we show the "constitutional navigation" of drones, i.e., encoding public aviation laws in an ASN for routing Unmanned Aerial Vehicles in uncertain environments.}, Keywords={Answer Set Programming, Deep Learning, Neuro-Symbolic AI, Large Language Models}, Crossref={https://github.com/ml-research/answersetnetworks}, year={2024}, @@ -21,7 +21,7 @@ @misc{skryagin2024asn Howbulished={arXiv preprint arXiv:2412.14814}, archivePrefix={arXiv}, primaryClass={cs.AI}, - url={https://arxiv.org/abs/2412.14814}, + url={https://arxiv.org/abs/2412.14814}, } @article{helff2024vlol,