Joho the Blogmachine learning Archives - Joho the Blog

December 17, 2017

[liveblog] Ulla Richardson on a game that teaches reading

I’m at the STEAM ed Finland conference in Jyväskylä where Ulla Richardson is going to talk about GraphoLearn, an adaptive learning method for learning to read.

NOTE: Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words. You are warned, people.


Ulla has been working on the Jyväskylä< Longitudinal Study of Dyslexia (JLD). Globally, one third of people can’t read or have poor reading skills. One fifth of Europe also. About 15% of children have learning disabilities.


One Issue: knowing which sound goes with which letters. GraphoLearn is a game to help students with this, developed by a multidisciplinary team. You learn a word by connecting a sound to a written letter. Then you can move to syllables and words. The game teaches by trial and error. If you get it wrong, it immediately tells you the correct sound. It uses a simple adaptive approach to select the wrong choices that are presented. The game aims at being entertaining, and motivates also with points and rewards. It’s a multi-modal system: visual and audio. It helps dyslexics by training them on the distinctions between sounds. Unlike human beings, it never displays any impatience.

It adapts to the user’s skill level, automatically assessing performance and aiming at at 80% accuracy so that it’s challenging but not too challenging.


13,000 players have played in Finland, and more in other languages. Ulla displays data that shows positive results among students who use GraphoLearn, including when teaching English where every letter has multiple pronunciations.


There are some difficulties analyzing the logs: there’s great variability in how kids play the game, how long they play, etc. There’s no background info on the students. [I missed some of this.] There’s an opportunity to come up with new ways to understand and analyze this data.


Q&A


Q: Your work is amazing. When I was learning English I could already read Finnish, so I made natural mispronunciations of ape, anarchist, etc. How do you cope with this?


A: Spoken and written English are like separate languages, especially if Finnish is your first language where each letter has only one pronunciation. You need a bigger unit to teach a language like English. That’s why we have the Rime approach where we show the letters in more context. [I may have gotten this wrong.]


Q: How hard is it to adapt the game to each language’s logic?


A: It’s hard.

Be the first to comment »

December 16, 2017

[liveblog] Mirka Saarela and Sanna Juutinen on analyzing education data

I’m at the STEAM ed Finland conference in Jyväskylä. Mirka Saarela and Sanna Juutinen are talking about their analysis of education data.

NOTE: Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words. You are warned, people.


There’s a triennial worldwide study by the OECD to assess students. Usually, people are only interested in its ranking of education by country. Finland does extremely well at this. This is surprising because Finland does not do particularly well in the factors that are taken to produce high quality educational systems. So Finnish ed has been studied extensively. PISA augments this analysis using learning analytics. (The US does at best average in the OECD ranking.)


Traditional research usually starts with the literature, develops a hypothesis, collects the data, and checks the result. PISA’s data mining approach starts with the data. “We want to find a needle in the haystack, but we don’t know what the needle looks like.” That is, they don’t know what type of pattern to look for.


Results of 2012 PISA: If you cluster all 24M students with their characteristics and attitudes without regard to their country you get clusters for Asia, developing world, Islamic, western countries. So, that maps well.


For Finland, the most salient factor seems to be its comprehensive school system that promotes equality and equity.

In 2015 for the first time there was a computerized test environment available. Most students used it. The logfile recorded how long students spent on a task and the number of activities (mouse clicks, etc.) as well as the score. They examined the Finnish log file to find student profiles, related to student’s strategies and knowledge. Their analysis found five different clusters. [I can’t read the slide from here. Sorry.] They are still studying what this tells us. (They purposefully have not yet factored in gender.)


Nov. 2017 results showed that girls did far better than boys. The test was done in a chat environment which might have been more familiar for the girls? Is the computerization of the tests affecting the results? Is the computerization of education affecting the results? More research is needed.


Q&A


Q: Does the clustering suggest interventions? E.g., “Slow down. Less clicking.”

A: [I couldn’t quite hear the answer, but I think the answer is that it needs more analysis. I think.]


Q: I work for ETS. Are the slides available?


A: Yes, but the research isn’t public yet.

Be the first to comment »

[liveblog] Harri Ketamo on micro-learning

I’m at the STEAM ed Finland conference in Jyväskylä. Harri Ketamo is giving a talk on “micro-learning.” He recently won a prestigious prize for the best new ideas in Finland. He is interested in the use of AI for learning.

NOTE: Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words. You are warned, people.

We don’t have enough good teachers globally, so we have to think about ed in new ways, Harri says. Can we use AI to bring good ed to everyone without hiring 200M new teachers globally? If we paid teachers equivalent to doctors and lawyers, we could hire those 200M. But we apparently not willing to do that.


One challenge: Career coaching. What do you want to study? Why? What are the skills you need? What do you need to know?


His company does natural language analysis — not word matches, but meaning. As an example he shows a shareholder agreement. Such agreements always have the same elements. After being trained on law, his company’s AI can create a map of the topic and analyze a block of text to see if it covers the legal requirements…the sort of work that a legal assistant does. For some standard agreements, we may soon not need lawyers, he predicts.


The system’s language model is a mess of words and relations. But if you zoom out from the map, the AI has clustered the concepts. At the Slush Sanghai conference, his AI could develop a list of the companies a customer might want to meet based on a text analysis of the companies’ web sites, etc. Likewise if your business is looking for help with a project.


Finland has a lot of public data about skills and openings. Universities’ curricula are publicly available.[Yay!] Unlike LinkedIn, all this data is public. Harri shows a map that displays the skills and competencies Finnish businesses want and the matching training offered by Finnish universities. The system can explore public information about a user and map that to available jobs and the training that is required and available for it. The available jobs are listed with relevancy expressed as a percentage. It can also look internationally to find matches.


The AI can also put together a course for a topic that a user needs. It can tell what the core concepts are by mining publications, courses, news, etc. The result is an interaction with a bot that talks with you in a Whatsapp like way. (See his paper “Agents and Analytics: A framework for educational data mining with games based learning”). It generates tests that show what a student needs to study if she gets a question wrong.


His newest project, in process: Libraries are the biggest collections of creative, educational material, so the AI ought to point people there. His software can find the common sources among courses and areas of study. It can discover the skills and competencies that materials can teach. This lets it cluster materials around degree programs. It can also generate micro-educational programs, curating a collection of readings.

His platform has an open an API. See Headai.

Q&A


Q: Have you done controlled experiments?


A: Yes. We’ve found that people get 20-40% better performance when our software is used in blended model, i.e., with a human teacher. It helps motivate people if they can see the areas they need to work on disappear over time.


Q: The sw only found male authors in the example you put up of automatically collated materials.


A: Small training set. Gender is not part of the metadata in Finland.


A: Don’t you worry that your system will exacerbate bias?


Q: Humans are biased. AI is a black box. We need to think about how to manage this


Q: [me] Are the topics generated from the content? Or do you start off with an ontology?


A: It creates its ontology out of the data.


Q: [me] Are you committing to make sure that the results of your AI do not reflect the built in biases?


A: Our news system on the Web presents a range of views. We need to think about how to do this for gender issues with the course software.

Be the first to comment »

December 15, 2017

[liveblog] Sonja Amadae on computational creativity

I’m at the STEAM ed Finland conference in Jyväskylä. Sonja Amadae at Swansea University (also currently at Helsinki U.) works on robotic ethics. She will argue in this talk that computers are algorithmic, that they only do what they’re programmed to do, that they don’t understand what they’re doing and they don’t feel human experience. AI is, she concludes, a tool.

NOTE: Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words. You are warned, people.


AI is like a human prosthetic that helps us walk. AI is an enhancement of human capabilities.


She will talk about about three cases.


Case 1: Generating a Rembrandt


A bank funded a project [cool site about it] to see what would happen if a computer had all of the data about Rembrandt’s portraits. They quantified the paintings: types, facial aspects including the size and distance of facial features, depth, contour, etc. They programmed the algorithm to create a portrait. The result was quite goood. People were pleased. Of course, it painted a white male. Is that creativity?


We are now recogizing the biases widespread in AI. E.g., “Biased algorithms are everywhere and one seeems to care” in MIT Tech Review by Will Knight. She also points to the time that Google mistakenly tagged black people as “gorillas.” So, we know there are limitations.


So, we fix the problem…and we end up with facial recognition systems so good that China can identify jaywalkers from surveillance cams, and then they post their images and names on large screens at the intersections.


Case 2: Forgery detection


The aim of one project was to detect forgeries. It was built on work done by Marits Michel van Dantzig in the 1950s. He looked at the brushstrokes on a painting; artists have signature brushstrokes. Each painting has on average 80,000 brushstrokes. A computer can compare a suspect painting’s brushstrokes with the legitimate brushstrokes of the artist. The result: the AI could identify forgeries 80% of the time from a single stroke.


Case 3: Computational creativity


She cites Wikipedia on Computational Creativity because she thinks it gets it roughly right:

Computational creativity (also known as artificial creativity, mechanical creativity, creative computing or creative computation) is a multidisciplinary endeavour that is located at the intersection of the fields of artificial intelligence, cognitive psychology, philosophy, and the arts.
The goal of computational creativity is to model, simulate or replicate creativity using a computer, to achieve one of several ends:[

  • To construct a program or computer capable of human-level creativity.

  • To better understand human creativity and to formulate an algorithmic perspective on creative behavior in humans.

  • To design programs that can enhance human creativity without necessarily being creative themselves.

She also quotes John McCarthy:

‘To ascribe certain beliefs, knowledge, free will, intentions, consciousness, abilities, or wants to a machine or computer program is legitimate when such an ascription expresses the same information about the machine that it expresses about a person.’


If you google “computational art” and you’ll see many pictures created computationally. [Or here.] Is it genuine creativity? What’s going on here?


We know that AI’s products can be taken as human. A poem created by AI won a poetry contest for humans. E.g., “A home transformed by the lightning the balanced alcoves smother” But the AI doesn’t know it’s making a poem.


Can AI make art? Well, art is in the eye of the beholder, so if you think it’s art, it’s art. But philosophically, we need to recall Turing-Church Computability which states that the computation “need not be intelligible to the one calculating.” The fact that computers can create works that look creative does not mean that the machines have the awareness required for creativity.


Can the operations of the brain be simulated on a computer? The Turing-Church statement says yes. But now we have computing so advanced that it’s unpredictable, probabilistic, and is beyond human capability. But the computations need not be artistic to the one computing.


Computability has limits:


1. Information and data are not meaning or knowledge.


2. Every single moment of existence is unique in the universe. Every single moment we see a unique aspect of the world. A Turing computer can’t see the outside world. It only has what it’s internal to it.


3. Human mind has existential experience.


4. The mind can reflect on itself.


5. Scott Aaronson says that humans can exercise free will and AI cannot, based on quantum theory. [Something about quantum free states.]


6.The universe has non-computable systems. Equilibrium paths?


“Aspect seeing” means that we can make a choice about how we see what we see. And each moment of each aspect is unique in time.


In SF, the SPCA uses a robot to chase away homeless people. Robots cannot exercise compassion.


Computers compute. Humans create. Creativity is not computable.


Q&A


Q: [me] Very interesting talk. What’s at stake in the question?


A: AI has had such a huge presence in our lives. There’s a power of thinking about rationality as computation. Gets best articulated in game theory. Can we conclude that this game theoretical rationality — the foundational understanding of rationality — is computable? Do human brings anything to the table? This leads to an argument for the obsolescence of the human. If we’re just computational, then we aren’t capable of any creativity. Or free will. That’s what’s ultimately at stake here.


Q: Can we create machines that are irrational, and have them bring a more human creativity?


A: There are many more types of rationality than game theory sort. E.g., we are rational in connection with one another working toward shared goals. The dichotomy between the rational and irrational is not sufficient.
TAGS:

Be the first to comment »

August 13, 2017

Machine learning cocktails

Inspired by fabulously wrong paint colors that Janelle Shane’s generated by running existing paint names through a machine learning system, and then by an hilarious experiment in dog breed names by my friend Matthew Battles, I decided to run some data through a beginner’s machine learning algorithm by karpathy.

I fed a list of cocktail names in as data to an unaltered copy of karpathy’s code. After several hundred thousand iterations, here’s a highly curated list of results:

  • French Connerini Mot
  • Freside
  • Rumibiipl
  • Freacher
  • Agtaitane
  • Black Silraian
  • Brack Rickwitr
  • Hang
  • boonihat
  • Tuxon
  • Bachutta B
  • My Faira
  • Blamaker
  • Salila and Tonic
  • Tequila Sou
  • Iriblon
  • Saradise
  • Ponch
  • Deiver
  • Plaltsica
  • Bounchat
  • Loner
  • Hullow
  • Keviy Corpse der
  • KreckFlirch 75
  • Favoyaloo
  • Black Ruskey
  • Avigorrer
  • Anian
  • Par’sHance
  • Salise
  • Tequila slondy
  • Corpee Appant
  • Coo Bogonhee
  • Coakey Cacarvib
  • Srizzd
  • Black Rosih
  • Cacalirr
  • Falay Mund
  • Frize
  • Rabgel
  • FomnFee After
  • Pegur
  • Missoadi Mangoy Rpey Cockty e
  • Banilatco
  • Zortenkare
  • Riscaporoc
  • Gin Choler Lady or Delilah
  • Bobbianch 75
  • Kir Roy Marnin Puter
  • Freake
  • Biaktee
  • Coske Slommer Roy Dog
  • Mo Kockey
  • Sane
  • Briney
  • Bubpeinker
  • Rustin Fington Lang T
  • Kiand Tea
  • Malmooo
  • Batidmi m
  • Pint Julep
  • Funktterchem
  • Gindy
  • Mod Brandy
  • Kkertina Blundy Coler Lady
  • Blue Lago’sil
  • Mnakesono Make
  • gizzle
  • Whimleez
  • Brand Corp Mook
  • Nixonkey
  • Plirrini
  • Oo Cog
  • Bloee Pluse
  • Kremlin Colone Pank
  • Slirroyane Hook
  • Lime Rim Swizzle
  • Ropsinianere
  • Blandy
  • Flinge
  • Daago
  • Tuefdequila Slandy
  • Stindy
  • Fizzy Mpllveloos
  • Bangelle Conkerish
  • Bnoo Bule Carge Rockai Ma
  • Biange Tupilang Volcano
  • Fluffy Crica
  • Frorc
  • Orandy Sour
  • The candy Dargr
  • SrackCande
  • The Kake
  • Brandy Monkliver
  • Jack Russian
  • Prince of Walo Moskeras
  • El Toro Loco Patyhoon
  • Rob Womb
  • Tom and Jurr Bumb
  • She Whescakawmbo Woake
  • Gidcapore Sling
  • Mys-Tal Conkey
  • Bocooman Irion anlis
  • Ange Cocktaipopa
  • Sex Roy
  • Ruby Dunch
  • Tergea Cacarino burp Komb
  • Ringadot
  • Manhatter
  • Bloo Wommer
  • Kremlin Lani Lady
  • Negronee Lince
  • Peady-Panky on the Beach

Then I added to the original list of cocktails a list of Western philosophers. After about 1.4 million iterations, here’s a curated list:

  • Wotticolus
  • Lobquidibet
  • Mores of Cunge
  • Ruck Velvet
  • Moscow Muáred
  • Elngexetas of Nissone
  • Johkey Bull
  • Zoo Haul
  • Paredo-fleKrpol
  • Whithetery Bacady Mallan
  • Greekeizer
  • Frellinki
  • Made orass
  • Wellis Cocota
  • Giued Cackey-Glaxion
  • Mary Slire
  • Robon Moot
  • Cock Vullon Dases
  • Loscorins of Velayzer
  • Adg Cock Volly
  • Flamanglavere Manettani
  • J.N. tust
  • Groscho Rob
  • Killiam of Orin
  • Fenck Viele Jeapl
  • Gin and Shittenteisg Bura
  • buzdinkor de Mar
  • J. Apinemberidera
  • Nickey Bull
  • Fishomiunr Slmester
  • Chimio de Cuckble Golley
  • Zoo b Revey Wiickes
  • P.O. Hewllan o
  • Hlack Rossey
  • Coolle Wilerbus
  • Paipirista Vico
  • Sadebuss of Nissone
  • Sexoo
  • Parodabo Blazmeg
  • Framidozshat
  • Almiud Iquineme
  • P.D. Sullarmus
  • Baamble Nogrsan
  • G.W.J. . Malley
  • Aphith Cart
  • C.G. Oudy Martine ram
  • Flickani
  • Postine Bland
  • Purch
  • Caul Potkey
  • J.O. de la Matha
  • Porel
  • Flickhaitey Colle
  • Bumbat
  • Mimonxo
  • Zozky Old the Sevila
  • Marenide Momben Coust Bomb
  • Barask’s Spacos Sasttin
  • Th mlug
  • Bloolllamand Royes
  • Hackey Sair
  • Nick Russonack
  • Fipple buck
  • G.W.F. Heer Lach Kemlse Male

Yes, we need not worry about human bartenders, cocktail designers, or philosophers being replaced by this particular algorithm. On the other hand, this is algorithm consists of a handful of lines of code and was applied blindly by a person dumber than it. Presumably SkyNet — or the next version of Microsoft Clippy — will be significantly more sophisticated than that.

Comments Off on Machine learning cocktails