Applied AI Pod

Episode highlights: * 01:00 - Conversational AI for the future of marketing and sales, focus on the real estate industry. * 04:00 - How Structurely works and what it solves. * 06:50 - Benefits to businesses utilizing AI within their companies. * 10:55 - The future of real estate by use of machine learning. * 16:10 - Creating a more promising future for AI as a tool for positive outcomes. E.g. Zillow. * 23:00 - Conversational AI's next big challenges. References: * Nate's LinkedIn profile * Nate's Twitter profile [https://twitter.com/whonatejoens?lang=en] * Structurely's Company Website [https://www.structurely.com/]

* 02:00 - Ada's performance, stories and metrics around. Size of the impact AI has in this space, as covered by Tradeshift. * 05:35 - Working with AI/ML teams. * 14:40 - Assessing how much data is needed for an AI project. * 18:45 - Data risks. * 24:25 - Is Agile good for AI teams? * 27:30 - How much does UX matter in e-Invoicing and ML/Data projects? * 36:35 - How can projects get derailed or fail? What should we watch out for. * 40:05 - Funny fails. * 41:50 - AI principles. References: * Lloyd's Linkedin Profile [https://www.linkedin.com/in/lloyd-humphreys-3b24009/] * Tradeshift's Ada technology [https://tradeshift.com/press/tradeshifts-next-level-ai-puts-payables-departments-in-control/] * Tradeshift's surpass of $1 trillion in transactions [https://www.businesswire.com/news/home/20210726005347/en/Tradeshift-Passes-1-Trillion-Transaction-Processing-Milestone] processed on their platform.

* 02:35 - Why hasn’t voice AI taken off already? * 22:50 - Can we fulfil an end to end new purchase naturally? * 32:20 - How can we resolve the disambiguation problem in NLU? * 37:20 - Context and memory perspectives. * 43:20 - How do we make conversations natural? References: * Dustin's VUX World Podcast [https://vux.world/podcast/] * Dustin's Linkedin profile [https://www.linkedin.com/in/dustincoates/] * Hannes' LinkedIn profile [https://www.linkedin.com/in/hannesheikinheimo/] * Speechly's Twitter profile [https://twitter.com/speechlyapi] * Speechly product search and checkout demo [https://www.youtube.com/watch?v=AlI47qnvip4] * Speechly's Interspeech Research Paper 2021 [https://www.isca-speech.org/archive/pdfs/interspeech_2021/pylkkonen21_interspeech.pdf]

* 01:15 - How does NLP work? * 04:05 - How do Transformer-based NLP models work? * 08:20 - How to look at unstructured data to take advantage of it more. * 12:00 - How to leverage ML to bring more to unstructured data? * 15:25 - Approach for low resources languages. * 23:25 - Word embeddings for common reasoning needs. * 26:55 - Techniques to follow to improve error and ambiguity in training data or for a model in general. * 30:10 - Are GPTs leading effort in the field in a wrong direction? * 34:15 -  Is DeepLearning the end of AI? * 37:20 - What are some good NLP metrics to watch? * 42:05 - How do we get past transactional queries to conversational queries? * 52:00 - Is the Turing test still relevant for NLP or has it become obsolete? References: * AI-Powered Search [https://www.manning.com/books/ai-powered-search]referenced in respect of text not being unstructured. * Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing [https://arxiv.org/abs/2107.13586] * Rethinking Search:Making Experts out of Dilettantes Common sense reasoning [https://www.researchgate.net/publication/351369052_Rethinking_Search_Making_Experts_out_of_Dilettantes] * TWIML AI podcast 518 with Yejin Choi [https://twimlai.com/social-commonsense-reasoning-with-yejin-choi/] * DARPA's Explainable AI Project [https://www.darpa.mil/program/explainable-artificial-intelligence] * EPITA [https://www.epita.fr/] is an engineering school in Paris. * Marc's LinkedIn profile [https://www.linkedin.com/in/marc-von-wyl-90639a2/].

* 12:50 - Is the Turing test still relevant? * 21:30 - Why it's important to use methodologies in AI projects and what are some best practices out there fit for AI projects. * 28:00 - Falsehoods of methodologies in AI projects. * 35:00 - Is Agile a good framework for AI/ML projects/products? * 40:10 - How can projects get derailed or fail if you don't have a plan in place. * 44:20 - The best compliment one can get after building an AI project or system. * 47:25 - Is DL the end of AI? References: * CPMAI methodology [https://www.cognilytica.com/cpmai/] * Cognilytica's Voice Assistant Benchmark 1.0 [https://medium.com/cognilytica/the-cognilytica-voice-assistant-benchmark-78455e747d46] and 2.0 [https://www.cognilytica.com/document/report-voice-assistant-benchmark-2-0-2019/] * AI Today podcast show with Alexandra Petrus as guest [https://www.cognilytica.com/2021/10/05/ai-today-podcast-insights-into-the-ai-startup-scene-interview-with-alexandra-petrus-host-of-the-applied-ai-pod/] * AI Today podcast show [https://podcasts.apple.com/us/podcast/ai-today-podcast-artificial-intelligence-insights-experts/id1279927057]