# Graph news # [Cold water for AI](https://readwise.io/reader/shared/01h8swnqyd2cednd1faxzn0cfz/) ![Growth rate of key innovations](https://i.imgur.com/8KvhluZ.png align="left") [Growth rate of key innovations](http://steveboese.squarespace.com/journal/2015/12/16/chart-of-the-day-the-technology-adoption-curve.html) You know what happens in cold water? Everything shrinks. > [Bing's market share hasn't grown at all.](https://www.zdnet.com/article/bings-search-market-share-fails-to-budge-despite-ai-push/) Bing's share of search It's still stuck at a lousy 3%. So [did traffic to ChatGPT](https://www.similarweb.com/blog/insights/ai-news/chatgpt-traffic-drops/) and other LLMs. What could be the possible reasons? * People realize it's **no magic bullet** - rather prone to hallucinations etc. All this despite OpenAI and OSS pushing out things like [AI agents](https://github.com/e2b-dev/awesome-ai-agents), [interactive fiction](https://github.com/mluogh/eastworld) and [multimodal models for healthcare](https://sites.research.google/med-palm/) * Again, a realization the **AI revolution won't take weeks, but months/years, if at all**. Despite that, people are losing jobs and [engaging in so-called 'sweatshops'](https://www.washingtonpost.com/world/2023/08/28/scale-ai-remotasks-philippines-artificial-intelligence/) - often it's the only possible income source. Yet still they are faced with platform bans, below-minimum-wage payouts and exploitation. It's a social issue I feel deserves outcrying. * Personally, I see it as a combination of trying to hastily **pump out deterministic outputs from a non-deterministic system** + a **one-armed bandit we're relentlessly pushing** the lever of, because each time we're getting an output which is approximately needed and plausible - yet all the time it's not the jackpot ![AI AI AI AI… AI!](https://i.imgur.com/yeN6RRV.jpg align="left") AI AI AI AI… AI! * AI raises and faces **legal**, **regulatory**, and **public** challenges - everyone's suddenly afraid of big corp using his/her personal data, scamming, yada yada. However, AI seems to excel (thanks to open source!) at deception, citing: * AI is now used in a series of [elaborate ransom scams](https://www.cnn.com/2023/04/29/us/ai-scam-calls-kidnapping-cec/index.html). * New AI bots [create malware on demand](https://decrypt.co/150899/wormgpt-fraudgpt-ai-hackers-phishing-malware). * Cheap AI music is used to replace human songs—not because it's better, but [because it's cheaper, and puts more power in the hands of technocrat platforms](https://www.honest-broker.com/p/twelve-brutal-truths-about-ai-music). * Students are [cheating with the aid of AI](https://www.theguardian.com/technology/2023/may/18/ai-cheating-teaching-chatgpt-students-college-university). * AI threatens to [disrupt the 2024 election with fake videos](https://www.wsj.com/articles/ais-rapid-growth-threatens-to-flood-2024-campaigns-with-fake-videos-dbd8144f). * Publications are misleading readers, who get [served up AI articles with little disclosure](https://futurism.com/the-byte/cnet-publishing-articles-by-ai). So, consumer demand is shrinking; with companies adopting the useful parts - the ones allowing them to run cheaper, extract more revenue - the usual. Remind something? ![Hmmm…](https://i.imgur.com/MmsYNJb.png align="left") [Hmmm…](https://www.ledgerinsights.com/gartner-blockchain-web3-hype-cycle/) Yeah, too for me. Are there any ways forward/around/inside? I believe yeah, one possibility lying in reasoning. # [Graph Machine Learning @ ICML 2023](https://towardsdatascience.com/graph-machine-learning-icml-2023-9b5e4306a1cc) Graphs + ML = ❤️ ![SE(3) Diffusion](https://github.com/jasonkyuyim/se3_diffusion/blob/master/media/denoising.gif?raw=true align="left") [SE(3) Diffusion](https://github.com/jasonkyuyim/se3_diffusion) Some of the papers' topics: * **New GNN architectures**: novel ideas of dynamically rewired message passing with delay and slow nodes for long-range tasks * **Generative Models**: e.g. diffusion models for graph generation (think Stable Diffusion that makes graphs) * **Molecules & proteins:** generating molecules, protein description (by its structure), predicting high resolution mass spectra (in 19 min instead of 126 hours - **x398 speedup**) * **Knowledge:** knowledge graph embedding ('packing'), discovering [both **unseen entities** and **relations**](https://openreview.net/forum?id=OoOpO0u4Xd) # [Neurosymbolic AI recordings & excerpts](https://twitter.com/asimunawar/status/1696850589299229130) ![By Knowable Magazine](https://i.imgur.com/rkOvVg9.png align="left") By Knowable Magazine ## Day 1. Crash Course in Neuro-Symbolic AI (Aug 29, 2023) [Introduction](https://www.youtube.com/watch?v=0WNlqqYdFi0) [Intro to Practical Knowledge Representation and Logic + State-of-the-Art Practical Reasoning and Meta-Logic (hour 2)](https://www.youtube.com/watch?v=SXrxBiSE_5o) [Logic with probabilities](https://www.youtube.com/watch?v=esbf06we2ss) [From Probabilistic Logics to Neurosymbolic AI](https://www.youtube.com/watch?v=H2iXZWuxz3w) [Reasoning with large language models](https://www.youtube.com/watch?v=jpMAwPZI6Ow) [Robust Logic: Past, Present and Future](https://www.youtube.com/watch?v=hW8pgWMjGrw) [Compositional generalization](https://www.youtube.com/watch?v=DQQCLzc5VfQ) [Panel](https://www.youtube.com/watch?v=APiAi3Y7VCA) ## Day 2: Diverse Approaches at the Research Frontier of Neurosymbolic AI (Aug 30, 2023) [Some extensions and applications of Robust Logic](https://www.youtube.com/watch?v=15EEcWztx9Y) [Deep Learning with Logical Requirements](https://www.youtube.com/watch?v=oB4dE7iUooA) [Neuro-vector symbolic architectures](https://www.youtube.com/watch?v=vG1oW5c9yvA) [AI can learn from data. But can it learn to reason?](https://www.youtube.com/watch?v=d5RVE830Pas) [Thinking fast and slow in AI planning](https://www.youtube.com/watch?v=8N73n56UlgM) [Utilizing knowledge in compositional generalization](https://www.youtube.com/watch?v=6G30ftk_CQA) [Causal abstraction for faithful, human-interpretable model explanations](https://www.youtube.com/watch?v=kU3CMrr5g1E) [Model-based ML: Towards causal reasoning in an AI Scientist](https://www.youtube.com/watch?v=tBOYLuLyK9Q) [Panel](https://www.youtube.com/watch?v=kRGURF54SFs) ## [NSAI Toolkit](https://ibm.github.io/neuro-symbolic-ai/toolkit) ![Source](https://i.imgur.com/ibaQj3y.png align="left") [Source](https://www.analyticsinsight.net/neurosymbolic-ai-know-about-the-next-ai-revolution/) A huge repository of IBM-developed neurosymbolic AI OSS, divided into 8 categories: * Logical Neural Network (LNN) * Natural language processing via reasoning (NLP) * Knowledge foundation (KF) * Learning with less (LwL) * Knowledge augmented sequential decision making (SDM) * Human in the loop (HIL) * Datasets and environments (DS) * Related advances (RA) Plus a nice [taxonomy of neurosymbolic systems](https://harshakokel.com/posts/neurosymbolic-systems/), which I won't include for brevity. --- Welcome to **Teleogenic/Boi Diaries**❣️ Other places I cross-post to: * [**Hashnode**](https://yawn.hashnode.dev/) * [**Medium**](https://baldr.medium.com/) * [**Telegram**](https://t.me/ohmyboi) * [**Twitter**](https://twitter.com/ZakharKogan) * [**LinkedIn**](https://www.linkedin.com/in/zakhar-kogan/)