More on the Turing Test: Loebner Prize, TTS, Hindi and other languages

Is conversational behavior shown by a naïve human being a good example of intelligent behavior? If a machine does as well in conversation as a human being to make you think that it is human, then many people will agree it is intelligent.

There is a lot of lore about the Turing Test, named after Alan Turing who defined it. The website http://www.cs.vu.nl/~jdruiter/c/index.html gives a lot of information on this.

The Loebner Prize

You may also wish to read http://www.loebner.net/Prizef/loebner-prize.html describing the $100,000 Loebner Prize for the first person to win the Turing Test by creating a suitable computer program. There is also an annual prize to be won by the best entry that year. The website mentioned above gives considerable information on the annual prizes awarded so far. Current technology has not yet demonstrated successful artificial intelligence at the level of winning the $100,000 Prize.

There is a very interesting discussion between Shieber [Lessons from a Restricted Turing Test] and
Loebner [ http://www.loebner.net/Prizef/In-response.html ]. Shieber has argued that if there had been a prize for a flying machine before the relevant science had been understood, people would have spent time trying to use springs to make flying machines. Loebner answers this by referring to Mozart’s backside!
To quote Loebner
“When Mozart rode to Vienna in 1781 he wrote complaining of the pain the mail coach inflicted on his backside. (Mozart in Vienna, V.Braunbehrens, trans T. Bell Grove Weidenfield, NY 1990, p 17). This was a result, I must suppose, in part from poor suspension of the coach. The study of elasticity, stress, and strain did not result in a swift and straight arrival at understanding. Suppose a concerted effort had been made, early on, to fly using springs. Perhaps the concepts of stress and strain would have been invented sooner, along with advances in spring technology that would have been a boon to humanity, and Mozart’s buttocks”.
Three cheers to Loebner! My project suggestion in this posting deals with an idea related in spirit. Many practical and valuable things can be achieved using the experience of working with programs designed to take the Turing Test. To continue with the analogy, instead of making a flying machine, one of these efforts could lead to an understanding of how to make properly sprung mail coaches.


Chatterbots

Before I start off with my proposal, let me mention Chatterbots. The Internet era has created “bots” which are programs which behave like robots; including the Chatterbots” which carry out conversations with humans. The website http://www.a-i.com/ gives you access to “Alan”, an interesting chatterbot.

Good, now we are ready to discuss my project suggestion. The proposal below involves extending the test idea to a spoken language context, not to win the Loebner prize, but to create a system, which could serve a variety of people in a limited way. I will argue that there are many attractive reasons to use a speech interface in the context of a Turing Machine. I do not believe that this is a magical solution to the challenge of making a machine show intelligent behavior, but I do believe that many students of artificial intelligence (AI) and computer science would consider this a promising way to advance the tools available to those who work on AI. The idea of using a speech interface is not new. You can find examples among winning programs from the annual Loebner prize contests that offer you a text-to-speech interface.

As I have mentioned earlier, there have been developments in search over the last ten years. Search engines dig out relevant information from millions of pages of text. I am going to deviate from the definition of the Turing Test, and will introduce, this time, the Ramani Test :=) This test will require a machine to respond to keyboarded questions in interesting and appropriate ways with spoken responses to impress listeners. It is deliberately a broad definition. Here is a relatively detailed proposal:


Marrying TTS capabilities in Hindi with the Turing Machine

The idea is to have humans communicate to a machine through a keyboard in one direction and through a test-to-speech system in the other direction. The focus would be on the machine being queried in one language, say in English, and responding in another language, say Hindi. The computer could do one of two following things, and the human participant would not be told which mode the computer uses at a given time.

1. The computer could simulate the conversational behavior of a person without significant school education – let us think of him/her as an adult coming to an adult literacy class in a village. The responses come through the TTS in the Hindi language (as an example). Questions can be asked in English and the simulated “person” who has only a very limited understanding of English would speak back in Hindi. An English version of what the simulated “he” says would be displayed on the screen so that the English-speaking human participant can understand what the answer is. The Hindi answer would be communicated through a speaker or a pair of earphones, to impress and amuse those who understand Hindi.
2. In the second mode of operation, the computer would display the question to a team of three, working somewhere out of sight of the person asking questions: a student (A) who reads the question and translates it an equivalent question in spoken Hindi. Another member of the team would be an actual member of an adult literacy class who answers these questions in Hindi. But the delegate would not hear the spoken answers directly. The student A would type the answer in Hindi using a suitable keyboard or a transliteration scheme, and the answer would go through the TTS to the delegate. (The simple transliteration scheme named ITRANS enables the use of a common Roman Script keyboard to input Hindi text). Simultaneously, another student, B, would convert the spoken Hindi answer to a typed-in English answer to be displayed to the original questioner.

You may ask why we use this elaborate method of going from the Hindi answer spoken by a human to its TTS “equivalent”. The idea is not to give away the nature of the “person” answering the question through the quality of the voice. This model ensures that the answers in both cases come only through the TTS.

3. Some delegates would be watching the three-person team at one end while others watch the delegate asking questions at another location. This is to demonstrate that we are not cheating!

4. The game is for the visitor to guess whether the answers came from the machine as in model (1) or from the real human being as in model (2). The system would randomly choose for each discussion session whether the respondent would be a real person, or the simulated person.

The big question is if the human participant would be able to distinguish the “simulated person” from the real one.

A socially valuable outcome of this work could very well be the wider understanding of two capabilities of uneducated adults: knowledge and communication ability.


Srinivasan Ramani Sept 26, 06


------------------

My Technorati Profile

Technorati Profile

Exploiting Search, Speech Recognition and TTS

August 27, 2006

There have been developments over the last 10 years that throw new light upon the challenge of communicating with a machine. Consider the great success of search engines in digging out the relevant out of millions of pages of text. This has made it possible for most of us to operate with a much higher degree of timely and relevant information than we would have had otherwise. This suggests that given adequate content in text form, search techniques could dig up and give us useful information in a user-friendly form, and there are lots of people who can benefit from such an interface.

Let us define a new test emulating the Turing Test, one called the Useful Talking Machine Test (UTMT). It will require a machine to respond to typed-in or spoken input in interesting and appropriate ways with spoken responses to help and impress listeners, on fairly broad topics defined by the test creator. It is deliberately a broad definition. Getting software/machines to win this test may be much easier than winning the Turing Test. Why?

The Turing Test leaves the questioner to turn the conversation on to anything and everything. The domain of discourse, and the expected performance criteria are undefined, except to say that if a human can cope with the expectation so should a machine. The UTMT, on the other hand, allows you two major comforts:

a) you can define the topic(s) on which the machine would take the test
b) it could use a large body of content (including all that is on the Web on this topic) to provide useful information to the person testing the system; most of the workers involved with the Turing test forget the value of large bodies of content.
c) The system could ride on the success of the search engines, searching through the stored body of text for paragraphs in which the search words appear.
d) School textbooks cover the common minimum knowledge that the average person at a given educational level has had access to. Hence, using the textbooks of ten school years could give us a certain level of base knowledge, which can be augmented by including selected books written at the right level and style.

However, TMT is not without its own challenges:

a) a search engine spews out findings recklessly, and expects to user to select the few relevant texts/visuals/recordings over the others
b) It does not specifically attempt to answer the user’s question by combining or rephrasing information discovered by search.

Consider asking why was Lincoln a great President of the US. I tried a search with the words {Lincoln, great president}. I looked into the first hit I got, searching for “great” and found this:

While the war raged, Lincoln also suffered great personal anguish over the death of his beloved son and the depressed mental condition of his wife.
(See http://www.americanpresident.org/history/abrahamlincoln/ )

So, let us define an experimental challenge:

Given multiple hits by a search engine working in response to the words in a natural language question, how can we synthesize a suitable reply to be given to the questioner as a set of selected, and connected sentences (or utterances)?

There are secondary issues, such as how do we filter, group and sequence the words in the question to ensure getting a meaningful reply. By grouping I mean using notations such as “great president”, or specifying that “Lincoln” and “great president” should occur in the same sentence. Filtering would eliminate the all-too-common words from the question while creating the search input. Such common words would bring out irrelevant hits of the search engine.

It would be useful to create a powerful context in which a system like this can be designed and in which the performance of the system can be judged. Let us examine two such contexts. They define frameworks in which the utility of the system proposed is emphasized, taking the focus away from comparisons with human performance. We started with the Turing Test, but have now come to define a significantly different test.

Context 1: An Assistant for Visually Handicapped Students

Here the computer would simulate a human assistant to visually handicapped students. The student would speak (or type in) a few words to indicate any need for information, and the assistant would “speak out” relevant information through a TTS system. Imagine that fifty textbooks are typed in (or read-in, if available in machine readable form) into the machine to provide the knowledge-base in the form of plain text. The answering technique would use a search engine to choose relevant paragraphs and then have them read-out using the TTS, possibly after appropriate transformation through an automatic “editing program”. The editing program could transform the sentences found by the search, to fit them to the question asked. It could eliminate unnecessary materials and duplications. It could also try to connect sentences using primarily syntactic transformations.

My own hunch is that selecting relevant sentences, ordering them in the right sequence, and not attempting to modify the sentences themselves would be the best approach. Wherever possible, the best solution would be to select a whole paragraph or a suitable part of such a paragraph and have it spoken out.

The text form is not the only one suitable for the knowledge-base. Someone might wish to consider representing the content of the books in some form of knowledge representation superior to that of plain text. This would, of course, necessitate the implementation of a search technique appropriate to the representation used. It would also raise the threshold of effort required to create the system, having to recast printed text into a suitable knowledge-representation. It would also necessitate a more elaborate exercise in producing comprehensible speech output.

The simplest assistant would merely select suitable hits from the search engine’s output, and them read out using the TTS. It could provide for some manual control, such as skipping what is being read-out when the user feels it is irrelevant, or having the read-out repeated when the user requests it. This approach would be useful if our focus is mostly utilitarian. On the other hand, the issue of synthesizing a suitable response from the raw output of the search engine is an important research challenge.

A Patient Information System:

This would be very similar to the Handicapped Student’s Assistant, but would work in the context of a medical patient’s helper. The system concerned could possibly have touch-screen input and stand in the lobby of a hospital. A patient, who is seeking information (on his/her way out, after meeting the doctor), could type in a question and have the machine speak out relevant information. For example, let us visualize some questions:

a) What are possible side effects of the medicine prescribed for me? (This assumes that the patient’s prescription information is made available to the system in some manner, without denying him privacy).
b) Can the regular use of calcium tablets cause bladder stones?
c) What causes TB?
d) What are the risks of having typhoid?

There are many places in the world where doctors do not have the time to deal with such questions from patients. That is where the helper would be of value.


Various other “skins”

There are various ways to clothe the basic scheme. One could make it look like an educational tool, offering a whole lot of useful information to a student at a school or in an adult literacy class. Examples: how to save money through a bank deposit, how to raise a bank loan, what are the dangers of pesticides, how to travel to a given destination, what happened at a given sports event, what a given person is famous for, etc. all restricted to the information contained in a set of books used in teaching at the school.

Steps Forward in this Project

The first step in addressing this problem would be to do a search for any earlier attempts to solve similar or related problems.

Discussion of this problem with others interested would have its advantages and disadvantages. The advantages are of getting a wider perspective from linguists, researchers in computer science areas such as human machine interaction and artificial intelligence, search specialists, school teachers, and teachers of the handicapped. The disadvantages are that it is always easier to think of many reasons to argue that something cannot be done! Canned wisdom cannot solve new problems, and often a little experimentation goes a long way.

The most important steps would be to critically assess what was tried and what was achieved, and to write about these.

A search for “natural language” “search engine output” produces several interesting hits. A particularly readable history of search engines is in
http://www.wiley.com/legacy/compbooks/sonnenreich/history.html


Srinivasan Ramani

Sea-going Robots

I was causally readig up on robots and found a set of nice links:

http://www.sciencedaily.com/releases/2005/03/050308134723.htm
does not talk in any detail about sea going robots, but has interesting news about MIT’s work on robotics.

http://news.bbc.co.uk/1/hi/sci/tech/286345.stm
reports work in Japan and talks about how much more energy efficient fish are compared to boats in regard to swimming.

http://peswiki.com/index.php/Directory:Wave_buoysdescribes how electricity can be generated using wave motion. However, they talk of big boys, oops big buoys, which can generate 250 KW of energy, when a shaft is moored to the ocean floor.

It is widely believed that life originated in water and then migrated to land. Similarly, I believe that significant robots would first succeed in the oceans. I do believe that water-borne robots with some limited form of “artificial intelligence” would soon be built. These are the grounds for my belief:

Navigation is easier, if you assume that your robot is too small to damage other craft in the ocean.

· Movement is easier.

· Autonomy and long life in the wild are made possible by three natural energy sources: wind, wave action, and sunlight.

· It is easy to set useful goals for ocean-going robots: get to such and such location, stay there and report weather information, oceanic measurements like wave activity. This justifies building them well.

Questions and Suggestions:

1. What prior work can you locate? What can you learn from it? Read up on digital buoys.

2. Who could you team up with?

3. Can you get academic credit for having fun building up this craft?

4. Read up about the basic theory of sailing. Can you visualize how the robot could deploy sails as necessary and be able to move?

5. Plan a secondary battery, which would have adequate working life. Plan a solar panel to provide electrical power to run a microprocessor onboard to give autonomous control, subject to a pre-loaded “script” which defines the purpose of the robot for the trip.

6. Can you design a way of generating electrical energy on board a small boat, utilizing wave action, to provide auxiliary power?

7. What should the craft look like? How stable would it be? Would it be sunk quickly in rough weather? Would you need to bail out water for preventing the craft sinking in such weather?

8. What materials would you use? What would be the design life of the craft? What are the challenges in keeping the craft intact and functioning?

9. What type of navigational instruments would you give it? GPS?

10. What type of communication capability would you build into the craft?

Separate the craft into sub-systems:

* The essential equipment to ensure autonomous behavior,
navigation & communication and measurements.

* The energy sources and secondary battery.

* The body that would enable floatation, waterproofing and structural rigidity.
Based on this, can you estimate the range of possible size and weight for this craft?

I will have more to say in the coming months on all this. Do contribute your comments and suggestions by posting them on this web page.

Srinivasan Ramani

(IN)HUMAN MACHINE INTERFACES

What you say to another human does not have to be perfect. There is a lot of redundancy in face-to-face communication. This helps the listener to check what you are saying for consistency and to ask for clarifications when necessary. But this comfort is not there when a website talks to millions of customers a year. I will describe a recent (bad) experience below. My focus will not be on the hassles I encountered except to illustrate what goes wrong. I feel that a good body of literature is available for helping us to design good websites and to test existing ones. I suggest that a good project could focus on what is wrong with a few heavily used websites, and provide suggestions to the companies concerned to improve them. A dissertation could cover the principles and techniques involved in evaluating websites and illustrate them with examples of existing problems with well-known websites.

Now to my experience. I recently flew a certain airline, which encourages passengers to do online check-in over the Internet. The biggest incentive offered is that you can save almost an hour of waiting in a long queue by doing your own check-in and printing of your own boarding card. But after the experience, I doubt the value of this incentive. I spent more time on the Internet than I would have spent in the queue. The hassles I faced are:

a) The rules were not clear. Failures were not explained properly. After three hours of trying and several phone calls they finally revealed to me that I could not do online check-in on my return journey though I had done it on the forward journey. I will discuss the reason later in the following paragraphs.
b) The website repeatedly told me “Our servers are not responding; please try later”, while it should have said, “You are not eligible for online check-in as you do not have an electronic ticket”. Online check-in is permitted by this airline to “cattle class” customers like me only within 24 hours of the departure. So to be told to try later causes great frustration with the deadline for check-in getting closer and closer.
c) I could have normally asked for an aisle seat over the phone. I tried to do this after waiting till online check-in was thrown open (24 hours before the flight) and then failing to succeed in this for an unknown reason. But the person on the toll-free number told me that a “general seat request” was not possible within 24 hours of the flight. Well, that is what they call a Catch-22! Neither could I select my own seat, nor could I ask any other human being.
Now to the hidden reason why the airline’s website tortured me for hours. The people I reached by calling the given toll-free number merely kept telling me to “try later”. Finally, I demanded to speak to the supervisor. The lady asked, “How would it help you? He would tell you the same thing as I am telling you. Bu I insisted, and she said, “Well, it might waste waiting for 10 minutes to speak to him”. I persisted and listened to some unwanted music for about 5 minutes, at the end of which got to speak to the supervisor. He now found out that I was not eligible for online check-in because they had converted my electronic ticket mid-way on the way forward to a printed ticket. Yes, they had done that. But why was every one ignorant of this rule, except the supervisor? And, why did the website not tell me this?
Let me go on a flash back now. On the way out from Bangalore, the flight was late in reaching the destination and I had to miss the connecting flight. They re-routed me to my destination to send me through some available flight. That is why I was downgraded from an electronic ticket holder to a mere mortal.
To close this part of the story, let me suggest that a few people should read literature on the subject. I will list a couple of them below.

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnsiteplan/html/improvingsiteusa.asp

http://www.library.yale.edu/~prowns/nebic/nebictalk.html

Now to the second part of my travails. After all this, I got on the flight home and again the flight was delayed due to some electrical fault that took time to fix. There was the risk of missing the connecting flight to Bangalore again. I talked to the staff at the counter who said that staff at the connecting location would take care of me. Again I demanded to see the manager, and I told him that the “staff at the connecting location had not done a thing to help me on the outward trip. I suggested that he should send a message ahead to ask that arriving passengers should be met on arrival by staff, who should help in negotiating the one-hour foot-and-bus journey between the arriving terminal and the departing terminal at the connecting location. He said that if there were a chance that I could make it to the connecting flight, the staff would inform me what to do. Well, what if there was no chance of my making it to the connecting flight? He did not say anything. I knew the answer from my experience. I would then do the one-hour trip to the “departing terminal”, find the flight gone and do the return one-hour trip to find some other flight going somewhere. Meanwhile my checked in baggage would travel as per its own free will!

What is the moral? It is not that we merely run lousy websites. We have lousy practices even without the Internet! When you deal with millions of people a year, it requires some good technology and a number of good design principles to ensure that they are not treated like cattle!

Ramani

Language Learning Games

Learning a foreign language can be fun. Even if you do not have a chance to travel abroad now, you can play a learning game on a computer, a personal digital assistant (PDA) or even on a cell phone. If you search for “language learning” & game, you find a number of commercial offerings. A few examples are:

http://www.transparent.com/games/

http://www.literacyconnections.com/0_052127737X.html

http://www.teflgames.com/why.html

However, there is always room for new games that help people learn a language. Developing such a game could be a great experience and offer you many benefits:

* Help you learn the language you are trying to teach. The teacher always learns more
than the taught!
* The experience can teach you about computer programming, particularly in
handling multi-media material and game development.
* The work can make you famous.
* The work might earn you credit in a computer programming course or in a foreign-
language course.

Many games use “fill up the blanks”, unscramble a word, and other similar techniques, to teach vocabulary. Very few exploit the multi-media capabilities of a computer or even that of a cell phone.

Imagine playing a game in which you meet “virtual people” who talk to you using displays of text or even “speaking out” recorded phrases. You can record short audio files containing one or two sentences, or even a short paragraph, in spoken form. A simple multi-media program can play these short audio files when showing relevant images – cartoons or photographs. The learner can view them, listen to the spoken words and respond in a suitable manner: answering a yes or no question, or typing in a word or number. He or she can even type in phrases that mean “you are welcome”, “thank you” or something like that. Talking to virtual people can be real fun, when you add a story line. Imagine you get off the plane, convert some money, get into a taxi, check into a hotel or meet grandparents or friends. You could create characters like a helpful taxi driver, a tough-talking policeman manning the taxi boarding area, a loving grandma and so on. Throw in a pretty girl, if you like! They can all be given suitable voices, by having a number of friends act out the scenarios.

Another mode of teaching could be top show images of objects (food, water, coffee, fruits, airports, post offices, a bus, a taxi, etc.) and have the learner type in the corresponding word. You should not forget the verbs. You can use images referring to run, sit, stand, go, come etc. You could play an audio file giving the spoken word with or without such image display, and expect the learner to type in the text form of the word. There is hardly any limit to ideas that such games can use.

The good thing about programming such games for use on cell phones is that people on a plane headed to a foreign country could use their time to learn the 30 essential phrases that could make a huge difference to their trip. Cell phone manufacturers offer software development kits (SDKs) on their websites to enable programming. This work is much easier than one expects at first sight.
Further it requires little equipment beyond a PC.

You might want to read about the Open Content movement (search the web for references), which creates resources for public use and puts them out on the Internet for free.

You could do a project in this area as an assignment for a foreign language course, or as a project for a science fair. Who knows? You may even sell such a game to a cell phone company!

Learn a bit about copyrights. You cannot merely lift images from other people’s websites and sell them off as a part of your game. You cannot put your creation out for free and then try to sell it.

Profit or no profit, you might end up teaching a few thousand people the elements of a language you love. It might be your mother tongue, or even Latin or Greek. Old myths and historical stories give you very interesting story lines for Latin and Greek.

A few relevant sources of information
--------------------------------------
http://en.wikipedia.org/wiki/Speech_recognition
http://www.speech.cs.cmu.edu/
http://www.faqs.org/docs/Linux-HOWTO/Speech-Recognition-HOWTO.html
http://support.microsoft.com/kb/306901/EN-US/
http://www.computertelephony.org/blog/


Have fun!

Srinivasan Ramani

Building an Intelligent Sailplane

The Intelligent Sailplane
March 19, 2006

This week, I read a very interesting item on detecting thermals on the website:

http://f4bscale.worldonline.co.uk/hand.htm

It talks about hand-launched model gliders and how to detect pockets of rising warm air into which you wish to launch your glider. It looks like young glider-builders get up to 60 seconds afloat for their model aircraft using these techniques. An interesting statement I read was that in the Northern Hemisphere, if you look up and see some evidence of a thermal, you will notice that the thermal is swirling clockwise! I have seen birds such as kites flying in lazy circles for half an hour or more at a stretch. I assume that they are riding a thermal, and fly along with the swirling wind and not against it! Stick your head out of the window and find out if they do clockwise circles.

The challenge

The question I wish to pose is this: Can we build a model sailplane controlled by a microprocessor on board, programmed to exploit any available thermals and stay up as long as possible?

This raises questions about identifying thermals. I will visualize one way of doing this and describe it. You may be able to think of better ones, or who knows, you may decide to build something like the sailplane I describe.

Detecting Thermals:

This could involve random exploration of airspace around the model craft while it is gliding. It should have sensors to detect increase in temperature as it flies into a warm column or bubble of air. This might mean a raise of 1 degree C over 5 to 10 meters of travel. Assuming that the glide speed is about 15 meters/second,
This would mean detecting 1 degree C change over about half a second.

Secondly, the craft should ideally have some sensor to give it information of its own altitude, and signs of any leftward acceleration. I assume that a glider being carried in a clockwise direction (as seen from the ground) will sense a constant leftward acceleration. The altitude sensor should tell the craft when it is going down fast and when it is being carried up by a thermal.

Can you use an altimeter, which works by sensing air pressure, for this purpose? I don’t know. Can we get inexpensive accelerometers to sense the leftward acceleration reliably? Again, I don’t know.

Assume that the craft is in a swirling column of air, being carried round in a circle of about 100 meters diameter, and that is gliding at a speed of about 15 Meters/second. What would be the leftward acceleration? How many “g”? It is best I leave this to you to calculate! The accelerometer you look for should be capable of sensing this reasonably well.

With all this gadgetry, I guess you might be able to build a computer-controlled sailplane that will exploit thermals.


Learning from Kite-like birds

I assume that they use a little bit of effort from their muscles to get the most out of the thermal. Understanding how they do this might help you use limited amounts of engine power (electrical drive, powered by solar cells on the wings?) to help your glider stay afloat longer. In any case, you have to develop techniques that will use input from the sensors to identify what should be done: Use the ailerons and rudder to make a gentle turn? Use the elevators to get the most of being carried around by the thermals? The secrets of the birds could tell us something about this.

Simulation:

Before you rush off to buy the parts for your project, I would suggest that you consult a physics teacher or a meteorologist. At least, do an Internet search to see if you can get information about the frequency and strength of thermals and when & where they occur. Then perhaps you could make a computer simulation of a cubic KM of air showing roughly the structure and statistical properties of your atmospheric model.

The next step would be to create a simulation of the controlling program you want to fly. Combined with the simulated atmospheric model, this simulated craft will let you figure out many issues, unrelated to your model-aircraft building skills! It may also teach a bit of atmospheric physics! The simulation itself might be a science fair project, or a research assignment.

I did an Internet search for “gliding simulators” and found two relevant sites:

http://shop.glidingshop.com and

http://www.answers.com/topic/glider

The first one describes a glider simulator which runs on a PC and the second one gives useful information.

Cheating

You might find a shortcut to the objective, by building a radio controlled model sailplane equipped with sensors that help you on the ground to use your head and steer the craft. It might be wise to do this first. This may be an alternative to simulation and hunting for atmospheric models. In fact it might help you earn a degree by doing research to create suitable atmospheric models. I suspect that useful atmospheric models for this purpose may not be available ready-made.

Have fun!

I dream of an autonomous model sailplane which would keep circling overhead my city and help us watch the traffic jams below! And, we do have plenty of the traffic jams!

Srinivasan Ramani

Is handwriting really accurate in identifying people?

There was a well-known article in a law journal in 1989, debunking the notion that a handwritten signature uniquely identifies the person who signs the document (Risinger, Denbeaux and Saks, 1989)

http://law.shu.edu/faculty/fulltime_faculty/risingmi/137UPaLRev731.pdf

A valuable introduction to this topic written much later can be found in

http://www2.warwick.ac.uk/fac/arts/ren/publications/lima/handwriting/forensic/

This latter article describes the safeguards prescribed by the US Supreme Court while evaluating the evidence given by Handwriting Experts. It also refers to a paper which gives the results of one study showing 7% errors in the case of experts and thirty eight percent in the case of non-experts. Note that people can be sent to jail ( or worse!) on the basis of testimony from "handwriting experts", or that your parents' property can go to someone else because of a disputed signature on a Will. I find 7% error rate among the experts very shocking. That is why we need to do more research and discuss this issue more in public.

But there is a widespread "conspiracy" to act as if this assumption need not be questioned. A lot of crime takes place in the world because this assumption is not correct. However, there are very few scientific publications that have tested this assumption. Can you think of a short research proposal saying how you can test this assumption? It would be excellent if you plan to do this research systematically and publish the result in a journal (printed journal or an online one). You might credit in school or college for doing this research. Further a teacher/advisor can guide you into planning the work to avoid weaknesses (experimental design).

The article I mentioned above says that handwriting verification of an individual is about as accurate as reading tea leaves to tell the future.

P. S. Please note there are two questions either of which you can address:

1. If a given signature is that of a known person whose sample signature is available? Yes or no?
This is relevant for instance when a bank verifies your signature on a check (cheque) to see
if it matches a "specimen signature" which you gave them while opening the account.

2. Given a signature is that of one of 30 people in a group, and given their sample signatures,
can you find out who the signatory is?
This is relevant for instance when the police try to identify a person out of a set of
suspects on the basis of a handwritten signature on a letter.
An additional question is if the identification works better when a whole handwritten letter
is available for comparison with handwriting on the supect's diary.

Suggestions:

a) Three relevant ideas are:

i) False positives: Cases in which the process being used identifies a signature as the one we
are looking for, while in fact it is not. This happens when the bank recognizes a forged
check as the real one!

ii) False negatives: Occurs for instnace when the bank refuses to pay against your
check saying the signature does not match the speciiimen signature.

iii) When you use comparison of handwriting to identify a criminal, how sure are you of the
result? 90%, OR 99%? or only 70%?

a) Avoid violating the law by forging someone else's signature as a part of the experiment. Think out how you can do something lawful, which allows you to test the hypothesis.

b) An experiment like this should involve at least 30 individuals to ensure that the result you get can be verified by someone else who repeats your published experiment with another group of volunteers.

c) It is not only the first publication that gets credit. A few following ones which test the repeatability of the finding will also be widely cited in future.

Have fun trying this out, and become the author of a research paper. As you know, they respect such things when they consider you for admission in college. Who knows, you might land up in Law School because of this work.

I hope teachers of statistics and experiment design would encourage their students to
try out these ideas.

Srinivasan Ramani