Beauty of Math

 The Beauty of Math

Quanta Magazine, 651K subscribers

"Students haven't been taught that math is discovery," says Richard Rusczyk, founder of Art of Problem Solving. "Math is a creative discipline—you're creating castles in the sky." Rusczyk has a vision for bringing “joyous, beautiful math” — and problem-solving — to classrooms everywhere. Read more at Quanta Magazine: https://www.quantamagazine.org 

Quanta Magazine is an editorially independent publication supported by the Simons Foundation.  ... ' 

From Game Playing to Protein Folding: Opening AI to Scientific Discovery

Excellent inspirational talk by Demis Hassabis on the breakthroughs deriving from Google company  DeepMind.  Which I have covered in this blog.    And also the ethics involved.  Largely non Technical and worth reading.  Note my brief touch on the challenge of protein folding.

Talk:  https://www.youtube.com/watch?v=jocWJiztxYA&t=4357s

Talk at the Institute for Ethics in AI Oxford

The past decade has seen incredible advances in the field of Artificial Intelligence (AI).

DeepMind has been in the vanguard of many of these big breakthroughs, pioneering the development of self-learning systems like AlphaGo, the first program to beat the world champion at the complex game of Go. Games have proven to be a great training ground for developing and testing AI algorithms, but the aim at DeepMind has always been to build general learning systems ultimately capable of solving important problems in the real world. I believe we are on the cusp of an exciting new era in science with AI poised to be a powerful tool for accelerating scientific discovery itself. We recently demonstrated this potential with our AlphaFold system, a solution to the 50-year grand challenge of protein structure prediction, culminating in the release of the most accurate and complete picture of the human proteome.

About Demis Hassabis: 

Demis Hassabis is the Founder and CEO of DeepMind, the world’s leading AI research company, and now an independent subsidiary of Alphabet. Founded in 2010, DeepMind has been at the forefront of the field ever since, producing landmark research breakthroughs. A chess and programming child prodigy, Demis coded the classic AI simulation game Theme Park aged 17. After graduating from Cambridge University in computer science with a double first, he founded pioneering videogames company Elixir Studios, and completed a PhD in cognitive neuroscience at UCL investigating memory and imagination processes. His work has been cited over 70,000 times and has featured in Science’s top 10 Breakthroughs of the Year on four separate occasions. He is a Fellow of the Royal Society, and the Royal Academy of Engineering. In 2017 he featured in the Time 100 list of most influential people, and in 2018 he was awarded a CBE.

About The Obert C. Tanner Lectures on Artificial Intelligence and Human Values:

The Tanner Lectures were established by the American scholar, industrialist, and philanthropist, Obert Clark Tanner. In creating the lectureships, Professor Tanner said: 'I hope these lectures will contribute to the intellectual and moral life of mankind. I see them simply as a search for a better understanding of human behaviour and human values. This understanding may be pursued for its own intrinsic worth, but it may also eventually have practical consequences for the quality of personal and social life.'  ...... 

Its Not Eureka

Isaac Asimov once wrote; "The most exciting phrase to hear in science, the one that heralds new discoveries,  is not 'Eureka'' (I found it!) but 'That's Funny' ... ' ".   ....    Understanding that quote is a real key to innovative thought. 

IBM Adds NLP to Watson Discovery

 In our early looks at Watson, precisely what we wanted to add to Discovery to make it useful.  For better links directly to business decision makers.   This cou d be key for novel uses, Show me some examples.

IBM to Add New Natural Language Processing Enhancements to Watson Discovery  in PRNewswire.

New planned features are designed to help business users quickly start applying AI to find more precise document insights with less training time and data science skills

Businesses in financial services, insurance and legal services turn to Watson Discovery to help automate processes and enhance customer care

NEWS PROVIDED BY IBM

ARMONK, N.Y., Nov. 10, 2021 /PRNewswire/ -- IBM (NYSE: IBM) today announced new natural language processing (NLP) enhancements planned for IBM Watson Discovery. These planned updates are designed to help business users in industries such as financial services, insurance and legal services enhance customer care and accelerate business processes by uncovering insights and synthesizing information from complex documents.

Businesses are increasingly turning to NLP and machine learning to help them comb through rising volumes of documents and data sets in a wide range of formats1. By applying AI to get document insights, business users can reduce research time and help their employees make more fact-driven decisions during complex, time sensitive tasks such as processing insurance claims, conducting financial analyses and reviewing legal agreements or contracts.

The new planned features that IBM announced today are designed to make it easier for Watson Discovery users to quickly customize the underlying NLP models on the unique language of their business. Stemming from NLP advancements developed by IBM Research, business users can train Watson Discovery to help read, understand and surface more precise insights from large sets of complex, industry-specific documents even if they don't have significant data science skills.

Pre-trained document structure understanding: Watson Discovery's Smart Document Understanding feature, available now in the Plus, Enterprise and Premium plans, includes a new pre-trained model that is designed to automatically understand the visual structure and layout of a document without additional training from a developer or data scientist. This helps users quickly find answers that were previously hidden or difficult to find like text in complex table structures or images.

Automatic text pattern detection: IBM has released a new advanced pattern creation feature in beta in the Plus, Premium and Enterprise plans that is designed to help users quickly identify business-specific text patterns within their documents. This is key for tasks like analyzing massive amounts of contracts or financial reports, which may report the same type of information, such as an increase or decrease in revenue, in different formats or using different phrases. Developed by IBM Research, it helps provide efficient ways of labeling data and training models. It is designed to start learning the underlying text patterns from as few as two examples and then refines the pattern based on user feedback. This helps users more rapidly train a model without manual and time-intensive tasks like defining rules and expressions.

Advanced NLP customization capabilities: Training NLP models to identify highly customized, business-specific words and phrases – for example insurance claim forms may include specific claim reasons or affected products – is a time-consuming task that requires significant data prep, labeling, and orchestration. Models trained on generic data sets often fail to retrieve the right information. With a new custom entity extractor feature, now available in beta for Watson Discovery Premium users, IBM is simplifying this process by reducing the effort for data prep, simplifying labeling with active learning and bulk annotation capabilities, and enabling simple model deployment that can accelerate training time.

The planned updates announced today are part of a pipeline of developments stemming from IBM Research. For example, answer finding was recently made generally available in Watson Discovery and Watson Assistant's Search Skill. It is designed to help busy professionals and customers identify the precise insights they need.  .... ' 

Material AI Invention and Discovery

New directions in new material discovery.  As it mentions, AI-driven molecular driven platform that is said to invent new molecular structure

IBM launches AI platform to discover new materials

Kyle Wiggers  @Kyle_L_Wiggers   in Venturebeat

IBM today announced the launch of the Molecule Generation Experience (MolGX), a cloud-based, AI-driven molecular design platform that automatically invents new molecular structures. MolGX, a part of IBM’s overarching strategy that aims to accelerate the discovery of new materials by 10 to 100 times, uncovers materials from the property targets of a given product.

The chemical sciences have made strides in the discovery of novel and useful materials over the past decades. For example, in the area of polymers, the recent development of thermoplastics has had an influence on applications ranging from new paints to clothing fibers. But while the discovery of new materials is the driving force in the expansion and improvement of industrial products, the vastness of chemical space likely exceeds the ability of human experts to explore even a fraction of it.  .. " 

Accelerating Discoveries with Data

Had not seen this effort yet, worth a look.

Accelerating data-driven discoveries

Life science companies use Paradigm4’s unique database management system to uncover new insights into human health.

Zach Winn | MIT News Office

As technologies like single-cell genomic sequencing, enhanced biomedical imaging, and medical “internet of things” devices proliferate, key discoveries about human health are increasingly found within vast troves of complex life science and health data.

But drawing meaningful conclusions from that data is a difficult problem that can involve piecing together different data types and manipulating huge data sets in response to varying scientific inquiries. The problem is as much about computer science as it is about other areas of science. That’s where Paradigm4 comes in.

The company, founded by Marilyn Matz SM ’80 and Turing Award winner and MIT Professor Michael Stonebraker, helps pharmaceutical companies, research institutes, and biotech companies turn data into insights.

It accomplishes this with a computational database management system that’s built from the ground up to host the diverse, multifaceted data at the frontiers of life science research. That includes data from sources like national biobanks, clinical trials, the medical internet of things, human cell atlases, medical images, environmental factors, and multi-omics, a field that includes the study of genomes, microbiomes, metabolomes, and more.

On top of the system’s unique architecture, the company has also built data preparation, metadata management, and analytics tools to help users find the important patterns and correlations lurking within all those numbers.

In many instances, customers are exploring data sets the founders say are too large and complex to be represented effectively by traditional database management systems.

“We’re keen to enable scientists and data scientists to do things they couldn’t do before by making it easier for them to deal with large-scale computation and machine-learning on diverse data,” Matz says. “We’re helping scientists and bioinformaticists with collaborative, reproducible research to ask and answer hard questions faster.”  .... " unique database management system to uncover new insights into human health.   ... "

AI and Legal Discovery

More developments for the textually and logic intense world of Law.  How much of legal tasks be replaced by systems? Following this.

Everlaw announces $62M Series C to continue modernizing legal discovery   By Ron Miller@ron_miller in TechCrunch

Everlaw is bringing modern data management, visualization and machine learning to eDiscovery, the process in which legal entities review large amounts of evidence to build a case. Today, the company announced a $62 million Series C investment.

CapitalG  (Alphabet’s growth equity investment fund) and Menlo Ventures led the round. Existing investors Andreessen Horowitz and K9 Ventures also participated. The startup has now raised $96 million, according to the company.

Everlaw  co-founder and CEO AJ Shankar says eDiscovery, which has been around for years, has become a classic big data problem. “We help legal professionals sift through huge volumes of evidence in lawsuits and investigations to find the smoking gun and the incriminating email,” Shankar told TechCrunch.

The software also helps teams of legal professionals work together and collaborate around this evidence. “Turns out that the law is incredibly collaborative, and we help these teams create a work product, and communicate and collaborate with each other in a system specifically built for the practice of law,” he explained.

He says this coordination is often done manually in spreadsheets with communication taking place via email, and even companies using legacy eDiscovery software are using systems designed in a time of lower data volumes.

The company offers a variety of tools to help humans locate the information they need to build a case. There is a search feature, of course, and data visualization tools including a timeline tool that helps pinpoint when key events happened. This can help lawyers direct researchers to find evidence within that critical period, greatly narrowing the focus of the search.

And the company also offers both supervised and unsupervised machine learning algorithms to help the team find specific bits of information. He acknowledges it will take human ingenuity working with the tool to find what you need. “No one’s handing you anything on a silver platter. It absolutely requires some detective work. It’s iterative and we have tools to help you with that,” he said.

Shankar stumbled into this area of technology when he was still a graduate student in computer science and a law firm came to his department looking for a technical expert. He ended up working with the firm for a couple of years, saw the kinds of technical challenges it faced, and decided to build some tooling to help..... "

Mining Scientific Applications by Analogies

This could also be done inside a company to look at publications like technical reports, and then also linking those to external publications as well.    Driven by specific goals as well.  The broad ability to use analogies usefully is a powerful thought.  The broad idea is one we looked at extensively, the tech is now here to do it.

AI analyzed 3.3 million scientific abstracts and discovered possible new materials  in MIT Technology Review

A new paper shows how natural-language processing can accelerate scientific discovery.
The context: Natural-language processing has seen major advancements in recent years, thanks to the development of unsupervised machine-learning techniques that are really good at capturing the relationships between words. They count how often and how closely words are used in relation to one another, and map those relationships in a three-dimensional vector space. The patterns can then be used to predict basic analogies like “man is to king as woman is to queen,” or to construct sentences and power things like autocomplete and other predictive text systems.

New application: A group of researchers have now used this technique to munch through 3.3 million scientific abstracts published between 1922 and 2018 in journals that would likely contain materials science research. The resulting word relationships captured fundamental knowledge within the field, including the structure of the periodic table and the way chemicals’ structures relate to their properties. The paper was published in Nature last week.

Because of the technique’s ability to compute analogies, it also found a number of chemical compounds that demonstrate properties similar to those of thermoelectric materials but have not been studied as such before. The researchers believe this could be a new way to mine existing scientific literature for previously unconsidered correlations and accelerate the advancement of research in a field.  .... "  ...

Discovering the Smart Home

Interesting thoughts here.   We are discovering.   After reading this I scanned my Alexa app and there were twenty of so 'smart' things.   There were about half as many in Google Home.  And I am still discovering, while they are still learning.   A few are unoperablefor various  reasons, the most common one is they are not connecting to the home Wifi architecture.    Another is that a device is not owned by the 'right' person, or IOT provider, and is awaiting updates.

Benedict Evans  in Andreesson Horowitz

Smart Home, Machine Learning and Discovery
April 04, 2019 In IoT, Machine Learning

‘Smart home’ today is in the same place as electric things in the home a generation or two ago: everyone will have some of these, but we’re working out which makes sense. Everyone got a toaster or a blender, but no-one got an electric can opener, and smart homes look the same. We’re in discovery mode.

My grandparents could have told you how many electric motors they owned. There was one in the car, one in the fridge, one in the vacuum cleaner, and they probably owned a dozen in total. Today we have no idea and it’s not a meaningful question, but we probably do know how many devices we own with a network connection. Again, our children and grandchildren will have no idea, and it won’t matter. 

In both of these cases, a wave of commodity components enabled a wave of product creation. The electrification of the home was enabled by cheap DC motors, heating elements and so on, and the current wave of ‘smart home’ devices is enabled by cheap and low power cameras, wifi chips, microphones and so on (mostly coming out of the smartphone supply chain).  

Equally, in both of these cases there’s a discovery phase: we may have all of these components but we still have to work out the right ways to combine then. Hence, people proposed all sorts of electric devices for the home, and we collectively worked out which made sense and where - everyone in Britain has a kettle, most people in America have a blender, and no-one has an electric can opener. 

The same is happening with ‘smart home’ now. Lots of ideas for products are being tried - some will be the kettles and some will be the can openers, and it will only be  obvious which in hindsight. Part of this process is also working out where the company value goes - which things are commodities from the existing manufacturers (oven companies, locks companies etc), which are commodities from Shenzhen, and which are opportunities for new company creation.  .... " 

Machine learning has a lot of the same questions: how do we combine these commodity components into products that make sense - both in the home and on platforms and smartphones? 

Part of the challenge of machine learning is not just working out what problems to solve but working out how to surface that to the user. Some of this may just be branding - we may need to say “this is ‘AI’” to set expectations (and also to lower them).  .... "

AI Enhanced Discovery and Optimization

A.I enhanced molecular discovery and optimization    By Flawnson Tong  in BlogUseJournal

We’re on the verge of a new scientific method, one that utilizes the power of artificial intelligence to accelerate the development of scientific discoveries.

“The amount of data being generated by experiments and simulations has given rise to the fourth paradigm of science over the last few years, which is data driven science, and it unifies the first three paradigms of theory, experiment, and computation/simulation.” — Ankit Agrawal of Northwestern University

Science can often be like a stick in the mud, hence the first two paradigms of theoretical and experimental science proved to be slow or two reasons:

lack of empirical evidence in theoretical knowledge (1st paradigm)
The abundance of human bias in experiments (2nd paradigm)

Computers however, changed the game for science, and we began to fully appreciate the accuracy and speed these incredible devices bring. The 3rd paradigm of computation and simulation was one of great promise, however limited by the technology of its time. This lead us to the 4th paradigm.

The 4th paradigm took advantage of a digital byproduct from the computation paradigm: data. Data science, while not an exact science, can lead to performance and accuracy so high that often times, they can identify new methods, equations, and ideas that we have yet to discover with science .... "

Nuance Project Pathfinder

Back to still the biggest problem for assistant AI.   How do we create and manage context rich, memory enabled conversations between people and machines?  Here the claim it is not best done by developing all possible scripts, but about 'intent discovery'.

Nuance's AI uses real interactions to make chat bots smarter in Engadget by By Rachel England, 

Project Pathfinder draws on human conversation, not service scripts.

 "... Instead of using specially trained "conversation designers" to create dialog patterns, Project Pathfinder leverages machine learning and AI to read existing transcripts of conversations between human service agents and customers, and creates a new workflow based on that data. Using what Nuance calls "intent discovery," the system is able to build a comprehensive conversation map to build a visual representation of all the different avenues its customer service conversations take -- from first question to each follow-up-- to reveal the best paths to resolution, as well as unknown problem areas. ... " 

More on Nuance Communications.
See their blog in Intent Discovery.

Discovering Skills in Virtual Assistants

Well done, analytical paper on the topic. Excerpt below and link to the full paper.  Recommended for people involved in using, designing and delivering skills for virtual assistants.

In the Communications of the ACM, November 2018 Vol 61 No. 11,   p 106-113

--------------------------------------------------------------------------
http://ryenwhite.com/papers/WhiteCACM2018a.pdf

Skill recommendations must be provided when users need them most, without being obtrusive or distracting.    By Ryen W. White,  Microsoft Research

Skill Discovery in Virtual Assistants

VIRTUAL ASSISTANTS LIKE Amazon Alexa, Microsoft Cortana, Google Assistant, and Apple Siri employ conversational experiences and language-understanding technologies to help users accomplish a range of tasks, from reminder creation to home automation. Voice is the primary means of engagement, and voice-activated assistants are growing in popularity; estimates as of June 2017 put the number of monthly active users of voice-based assistant devices in the U.S. at 36 million.a  Many are “headless” devices that lack displays. Smart speakers (such as Amazon Echo and Google Home) are among the most popular devices in this category.

Speakers are tethered to one location, but there are other settings where voice-activated assistants can be helpful, including automobiles (such as for suggesting  (such as for providing private notifications and suggestions18). Virtual assistant capabilities are commonly called “skills.” Skill functionality ranges from basic (such as timers, jokes, and reminders) to more advanced (such as music playback, calendar management, and home automation). 

Assistant skillsets include both first-party skills and third-party skills. First-party skills comprise the aforementioned basic skill functionality found in many assistants, as well as skills that leverage assistant providers’ strengths in such areas as electronic commerce (Amazon Alexa), productivity (Microsoft Cortana), and search (Google Assistant). All major assistants also provide development kits that empower third-party developers to create their own skills for inclusion. Skills can be invoked independently, linked together within a single voice command to invoke a preprogrammed routine, or in a sequence of related skills arranged as required for complex task completion. Despite the significant value virtual assistants can offer, discovery of their capabilities remains   a challenge  ... "

Commercializing Discovery

 Events yes, and needed process also.

Most Science Startups Fail. Here's Why
We need to get a lot better at bridging that gap between discovery and commercialization  By Greg Satell in Inc.

" ... The gap between discovery and commercialization is so notorious and fraught with danger that it's been unaffectionately called the "Valley of Death." Part of the problem is that you can't really commercialize a discovery, you can only commercialize a product and those are two very different things.

The truth is that innovation is never a single event, but a process of discovery, engineering and transformation. After something like graphene is discovered in the lab, it needs to be engineered into a useful product and then it has to gain adoption by winning customers in the marketplace. Those three things almost never happen in the same place. ... "