Tristan Purdy Colangelo

                                                                                                Psy 322

                                                                                                April 20, 2003

Proposal

            For my research project I would like to study the history of technology as educational devices.  Although not technically a machine by our standards the chalkboard created enormous disturbance within classroom as teachers had no idea how to use this tool to help shape and deliver lessons.  Throughout history with the overhead projectors and computers there has been much talk about how much these devices could potentially revolutionize the education field.  Most of these technological advantages have been met with criticism because they do not seem to meet their expectations.  I would like to explore the history of these devices and how they have created/diminished educational opportunities in the classroom.  I would assess which of these devices have made a long lasting impression in classrooms and which ones have made a quick exit from the scene.  I would like to try and explore why these devices have either been useful or have been useless. 

***************************

 

 

Mati Chessin

Ben Holskin

PSY/ORF 322

4/20/03

 

Final Project Proposal

 

Topic:  Why is it so hard to program VCRs. 

 

 

Advisors(s):  Professor Cooper

 

Objective:  We hope to determine what makes VCRs so hard to program.  From our findings, we will propose better ways to set up a VCR. 

 

 

Process:  We will examine common VCR interfaces, controls, inputs, outputs and hardware to isolate the basic principles necessary for VCR programming and how they relate to human processing.  We will also examine products that are designed to help program VCRs such as the VCR Co-Pilot to identify what elements facilitate the process.  If necessary, we will also conduct surveys to get human feedback on the subject. When possible, we will try to use any research already done on this subject.

 

 

Hypothesis:  We believe that most of the problems associated with programming VCRs centers around the connection between the VCR and the TV.  For many VCRs, there are straightforward menus to help set the time and record shows, but in order to get to those menus, there are a plethora of steps necessary to put the TV in the proper mode.  We suspect that another major problem concerns the differences from one VCR to the next.  So while a person may learn how to program one VCR, those skills may not translate well to all VCRs.  We also suspect that the difficulty associated with VCRs has led to VCR replacements like TiVo

******************88.

 

GENDER DIFFERENCES IN WEBSITE USER-INTERFACE

 

Matt Hawrilenko

 

Emily Kroshus

 

Advisor:  Professor Cooper (hopefully?)

 

Purpose

           

The purpose of our project is to examine how web developers utilize knowledge of gender differences when designing interfaces for typically gender-specific websites.  We will also examine the extent to which supposedly non-gendered websites incorporate elements of gendered design.

 

Process

 

We will examine websites that are explicitly tailored to gender-specific audiences (Maxim, Cosmo), websites that are supposedly gender neutral.    We will construct a list (based on class and outside readings) of what we consider theoretically successful design-interfaces in websites geared towards males and females and examine websites to see what type of designs are prevalent.  We will then analyze our findings based on what we expected to find and what surprised us and attempt to explain why some actuality deviated from our theoretical predictions.

 

Findings

 

We expect to find differences in techniques employed to create gender specific websites.  In the websites that are non-gender specific, we expect to find at least some embedded gender bias reflected in the interface design.

 

 

Is Humanity Ready for a Cashless Society?

 

Ryan Wells and Matt Wolf

Advised by Professor Kornhauser

 

We plan to study the question of whether or not people are ready to live in a society where all financial transactions are done electronically ie online purchasing, online bill paying etc.  Although there are undoubtedly advantages to a cashless society, it is possible that costs of implementation and people's reluctance to embrace such a system may outweigh these benefits.  In a human behavior study, we plan to address questions like what aspects of e-commere give people anxiety (security concerns etc.), how different groups of people will handle the transition (gender, age, computer experience, income etc.).  We hope that a preliminary search for literature on these topics will lead to additional issues which we can address.

 

 

Jason White

Candace Hamilton

Final Project Description

PSY/ORF 322:

Human – Machine Interaction

21 April 2003

 

Machines that aid humans in the decision making process are in increasing demand.  Because, as Professor Johnson-Laird pointed out, humans are particularly poor at logic and probability, machines can often serve to better make decisions than people.  While there are some emotional parameters for which machines can not account, these same parameters can also obscure a human’s ability to make the best decision. 

For example, at Princeton many students agonize over the issue of deciding which eating club to join.  There are a number of factors that can affect how a student will end up enjoying the environment of any one eating club.  Some of the factors are emotional and others are more pragmatic.  We are going to quantify each the most important of these parameters in an attempt to design a program that assists people in deciding what eating club they should join.  After building the model, we will run a series of tests on a small sample of students.  Through this brief testing, we will assess what eating club each person should have joined according to our model, what eating club the individual actually did join, and whether the individual, in retrospect, feels that eating club chosen by our model would have better suited them.  Clearly, there are a number of biases associated with the testing procedure.  How one can actually determine if they would have had a better or worse experience in a different eating club is extremely subjective.  However, for the purposes of acquiring some summary data, the information will be helpful. 

In the end, we will be able to present a study that draws on lectures from Professor Johnson-Laird and Professor Kornhauser, most heavily.  In addition, to the program which will assist students in selecting the correct eating club, we will have data that will examine the extent to which human-machine interaction with respect to eating club selection is limiting. 

**

 

Julie Kestenman and John Andrews

                                                                                    Psy/Orf 322

                                                                                    Project Proposal

           

 

Religiosity and Remote Random Processes

 

 

For our project, we want to modify the “paintings” PEAR lab experiment. In each run, the computer superimposes two pictures on top of one another, and then a random generator chooses which painting’s pixel should be displayed. During the tour of the lab, Brenda Dunne said that paintings with religious connotations seemed to show up more frequently than those without. We want to investigate this idea more closely.

 

Experiment 1:  Testing the effects of the spirituality of the viewer (X) on the outcome.  We will give the computer the choice of paintings between one of a religious and one of a secular theme, both by the same Renaissance master. (Y, the effect of the spirituality of the artist on the outcome, is presumably thus controlled.)

 

Experiment 2:  Testing the effects of Y on the outcome, controlling X.  We will give the computer the choice between an abstract work by a religious artist, and an abstract work generated by a (secular) computer.  (Presumably the viewer cannot distinguish the human-generated or the religious if both are abstract, and X is controlled.)

 

Experiment 3:  Testing the effects of X and Y on the outcome.  We will give the computer a choice between a representational religious work by a Renaissance master and an abstract work by a computer.

 

 

Following each participant’s run, we want to give each person a short survey asking them what their religion is, how religious they are, their perceptions about the experiment, and their methodology of trying to make one painting show up more than the other. Using this survey information coupled with the results of the experiment, we hope to find some sort of answer to our hypothesis.

 

Reference used will be Two Decades of PEAR, particularly Dunne and Jahn’s “Experiments in Remote Human/Machine Interaction,” pp 311-454.  These results suggest that we may wish to confine our experiment to females.

 

Brian Chang

Edward Cong

Project Advisor: Professor Cooper

 

The Long-term Consequences and Effectiveness of Educational Computer Games

 

We wish to study the long-term effects of educational computer games on children that played these games about 10-15 years ago, when schools first began using computers as educational tools.  By surveying college students (ages 18-24), we can find out directly from those who played the games what the effects were, now that they are older and better able to understand how the those games were designed teach them.  For example, the survey would discover what games were helpful in learning a certain subject, motivating an interest in certain subject, or games that were just fun but didn't help much in learning.

 

Example survey questions:

(1) What educational computer games did you play in school (elementary through high school)?

(2) What were each of these games supposed to teach you?

(3) Which games did you enjoy playing?

(4) Which games helped you learn, and what specifically did you learn?

(5) Which games helped motivate your interest in a certain subject?

(6) Which games were fun, but did not really help your education in any way?

 

From the results, we can analyze gender differences regarding the interest, motivation, and effectiveness of these games.  We anticipate that most college students will remember pretty accurately the games they played in school.  Also, they will probably have found those games fun and interesting to play, since most of those games included intrinsic motivational properties like those found in the readings.  However, probably many of the games will have been only that – fun – and not much use in the way of actual teaching.  We suspect that the games were more likely helpful in a way that motivated an interest in a subject, and motivated the student to learn more about it, rather than actually teaching the student the subject.  Hopefully, we will be able to determine how to optimally use computer games in schools.  Based on the findings, perhaps the best way may be to separate games in to purely motivational games and learning games.

 

 

Sarah Molouki

Project Title: The Internet: Tool or Trap?

Advisor/Reader: Professor Cooper

 

Objective:

The project will be a study of the motivational effects (similar to those discussed by Lepper and Malone in the study of educational computer games, e.g. control, fantasy, randomness, and others) which are most pertinent to the domain of internet usage for both educational and entertainment purposes. These factors will be used to make a distinction between internet tools and their non-online equivalents (e.g. online database vs. standard library, entertainment websites vs. TV or magazines) in order to try to discover what it is about the medium of the internet itself which makes it so intrinsically appealing to so many people. In particular, what are the motivational factors which can cause the internet can become too appealing in many cases, leading to problems such as overuse and even “internet addiction?”

Process:

            A background review of relevant literature will be conducted in order to identify the motivational factors which would be likely contribute to the popularity of internet usage. This will likely be followed up by an informal survey to see if people’s own assessments of their motivations for internet use in fact agree with these theoretical suggestions. In addition, some of the sites which people have found to be most appealing and/or addicting will be examined in order to attempt to identify which characteristics they share in common (or in contrast) with each other and/or with other motivationally appealing or addicting non-internet activities (TV, games, even gambling).

Expected findings:

            We will expect to find that the vast appeal of the internet will likely be due to the presence of many of the common motivational factors identified in the literature, such as randomness, control, and fantasy, which are present or seem to be present in the format of the internet media to a degree that they are not present in many similar non-online activities. This knowledge will hopefully lead to greater awareness of ways in which the internet can be effectively used, as well as serving as a warning for the potentially negative allure of the internet.

 

 

Aaron Zimmerman                                                                                            4/21/03

PSY/ORF 322

Final Project Topics

 

Note: Topics NOT listed in any particular order.

            - As of now I am most interested in #1

 

Advisor: Kornhauser

#1: Control Interface: Creating a Princeton WWW page for the people that would like to use it the most: The students (undergraduate focus)

-         We know students use their computers very frequently to gather information.

o       Bring the together as much information as possible in order to create a more efficient user experience

-         Research the idea of creating a student site: www.princeton.edu/students

§         Let students be able to enter a customized Princeton site, built by them and for them.

§         Conduct survey to find most used services.

§         Student specific content: helpful for daily life

·        Displayed on site: pretty standard

o       Weather, Movies, Headlines

o       Webmail / Blackboard / directory input fields

·        Links to other sites: standard + customizable

o       SCG, athletics, student agencies…

Advisor: Cooper/Harman?

#1: Implications of technology on moral values.

-         MP3 proliferation / file sharing

-         Informality and ease of emails

-         Pornography

-         Cost of bandwidth

-         Hacking

 

Advisor: Cooper/Harman?

#2: Implications of Instant Messenger on college campuses (any beyond)

-         Primary means of communication?

-         Away messages: do they now serve a purpose, could they?

-         Time consumption / distraction

 

 

 

Michelle Buffum

PSY/ORF 322

Project Proposal

 

 

Professor Cooper has agreed to advise me in the following project.

I am thinking of doing an online survey to be sent to undergraduate students. I have come across some people who bring their laptop with them everywhere they go and use it for everything from email and calendar to weatherman and jukebox and others who never get past the bare necessities of Microsoft word, email, and solitaire. The topic of dependency on machines has come up in class a few times lately and I am interested in gauging how addicted to computers students are. Perhaps I'll look at this attachment with respect to demographics (whether or not students from rural places are less attached to computers than those from more urban reaches) and experience with computer games (either in the classroom or on their own - perhaps this causes them to perceive their computer differently). I want to find out how addicted to computers people are, and what possible explanations might be.

 

 

Psy/Orf 322 – Human-Machine Interaction                                                                   4/21/03

 

Final Project Proposal

 

Title                          Game Theory and Human Decision-Making

Students                Greg Prisament and Nicole Di Lello

Faculty Advisor            Philip Johnson-Laird

Abstract

            Our final project in Psy/Orf 322 will look at human decision-making in games of strategy.  Game Theory provides mathematically “correct” solutions to games of strategy (such as the Prisoner’s Dilemma).  We will test humans to see if they make the decisions that Game Theory says they should.  From this we will attempt to discover how humans model games of strategy in their mind.

            To perform the test we will give a short survey (15-30 minutes) to a group of 15-25 participants.  The survey will consist of several games - either in matrix form or phrased as conditionals.  The participant must decide if he would play the game.  If he decides to play, he must circle his decision.  Perhaps we will also require the participant to explain his reasoning.  To provide incentive we will give the participant who makes the best decisions a small cash reward.

We have not decided on the exact details of the survey or what it will test.  A possible variable is feedback (whether the player is told the outcome of the game or not.)  We would compare human behavior in games with feedback to their behavior in games without feedback.  We believe humans will be more willing to play the games with feedback.  Also, they will be more concerned with their opponent’s outcomes in feedback games than in non-feedback games.

            We anticipate that humans will not always make the Game Theoretic “best” decision.  To explain the discrepancy we hope to find evidence for a mental model of games that differs from Game Theory in two ways:

        It is computationally less expensive (i.e. heuristically based).

        There are factors that are not captured by the mathematics of decision-making, such as risk aversion/predilection and a notion of “fairness.”

 

Our final project will include an overview of previous work on this subject, a detailed explanation of our experiment, an analysis of our findings, a proposed mental model, and a case for how machines can help humans make better decisions.

 

 

Title:  Improving Reasoning with Frequency and Chances Formats

Meg Gerbasi & Rebecca Gillespie

Professor Johnson-Laird

Abstract

The objective of this project is to determine if performance on probability format (.70) reasoning problems can be improved by doing frequency format problems (70 people) or chances format problems (7 out of 10).  In order to test this, participants will be given 3 problems to solve.  The problems will be identical except for variation in whether the numbers are given in terms of frequency, chances, or probability.  The first 2 problems will vary, such that 1/3 of the participants will do 2 frequency problems, 1/3 will do 2 chances problems, and 1/3 will do 2 probability problems (this is the control group).  Then all participants will do an identical 3rd problem in probability format.  The participants will do the work to solve the problems on paper given so that their strategy can be examined.  After completing the problems, participants will be asked to describe any strategy they had for solving the problems and for background information such as their gender, class year, and major.  Performance on the final problem will be the primary dependent measure.  Participants will be taken from students in the student center or in a residential college and will be given pizza for their efforts.  We anticipate that doing frequency format problems first should help participants develop a strategy and solve probability problems correctly later.  Chances format should also help, but since it is slightly more complicated than frequency format, the improvement effects may be smaller.  Participants in both the frequency and chances conditions should be better at the probability problem than participants in the control condition who only do problems in probability format.  The project should help to determine what learning techniques may be useful in improving reasoning and developing personal strategies for solving reasoning problems.   

 

Mari Kobayashi
PSY 322 / ORF 322
Final Project Proposal
4/21/03
Final Project Description
Title:        The Effectiveness of Technology in Education
Name:         Mari Kobayashi
Advisor: Professor Cooper
Abstract:
Objective:  
To assess the use of technology in K-12 education by evaluating its effects thus far, and further potentials in the future.
Process:     
First, I will research published articles and studies done on technology in education from both educational and technological perspectives. Next, I will evaluate the positive and negative effects of technology in order to analyze its importance and benefits to education. I will also survey students of various ages (K-12/college students) about their experience of computers uses in secondary and primary education to be included in the evaluation.
Anticipated Findings:
While the debate for and against the use of technology in education continues, the benefits and further potentials with continually developing technology outweigh the costs and time required to establish a solid foundation for sources and technologically trained and competent teachers to enhance education. This hypothesis calls for profitable opportunity for the software companies to develop and market educational tools, while creating greater learning experience to improve the quality of education.
 
 

PSY 322                                                                                                     Emmanuell Murray

Project Proposal                                                                                                April 21, 2003

 

Fear and the Human Processor - Kornhauser

 

Under normal conditions, simple reaction time is about tp + tc + tm = 240 ms.  Math problems can take considerably longer, but the time they take is still a function of these three time constants.  I propose that time to solve a problem should decrease for a situation in which fear of an undesired outcome is introduced.  I intend to test this hypothesis by creating an experiment in which people participate in two studies: one for control and one where fear is introduced.  In the control, people are asked to complete three addition problems.  This tests their initial ability to complete problems under normal circumstances.  In the fear test, the same people are asked to complete three problems of similar difficulty, but during the first problem a loud beep will occur after an amount of time less than the subject’s initial response time.  The subject will be told that he took too long to answer the question, and that he must continue to the next.  In order to avoid the beep, they should finish the problems faster than their control time.  The beep should only be used during the first problem, lest they get too used to the sound so that the fear aspect disappears.

 
Gender Differences and the Evolution of Computer Games
Eddie Chou (echou@princeton.edu)
Greg Hon (ghon@princeton.edu
Professor Cooper
 
Our research project is based upon Professor Cooper's lectures on the
attraction of computers, especially games, and existing gender gaps involved
in multiple aspects of these and related themes.  After establishing a
difference between genders towards computer gaming, we would like to apply
this knowledge to examine the evolution of popular game series, such as
Super Mario Bros., Final Fantasy, the Legend of Zelda, etc.  We hope to be
able to find relations between different aspects of the game, its
attractiveness to the two genders, and the popularity of the game.
 
 
 
Prof. A. Kornhauser-
 
For the PSY 322 Final Project, I would like to be paired with Mariam Vazquez.  We have discussed our project, and we would like to work with Professor Cooper.  As such, our project is tentatively titled "Educational Applications of Computer Technology in terms of Gender Differences".
Our project focuses on how the introduction of computers into classrooms raises questions concerning the differences on the existing gender gaps in computing and in education.  In order to explore this topic, we will approach it from three different dimensions.  The first is the current state of gender differences in education.  The second being the current state of gender differences in computing, and the third is the effects of computers in the classroom (i.e., Prof. Cooper's assigned readings).  Our project will then explore how these three topics interrelate, and we will make educated predictions about the effects computers will have on the gender gap for computing and education in the future based on our research.  We will base our project on Prof. Cooper's findings (i.e., the articles we've read in class), on research as recommended by Prof. Cooper, as well as research that my partner and I will find outside of class.
As Miriam and I have differing opinions on how to interpret such research for our predictions (as to whether or not the gender gap will decrease), we may present our project in such a way as to say "based on our research, I believe such-and-such, whereas my partner does not and this is why...".  But this may change upon consulting Prof. Cooper on Monday.  As this is more of a "library thesis", we have found it difficult to develop an "exciting" visual presentation aside from the usual tables and graphs.  However, we believe that we have a solid project which we can present at the Symposium on May 12th.
Due to time constraints involving JP's and other class-work, Mariam and I have not had the chance to contact Prof. Cooper as to the extent of our tentative project outline.  We will try to address this oversight on Monday, after class.
 
-Josephine L. Jim
PSY, 2004
 
 

 

 

 

 

GARRET WESTON      &       TIM RELEFORD

 

PSY/ORF 322 Final Project

 

 

“Mr. Frog Goes Riding: To Move Or Not To Move”

 

 

Advisor(s):  R.G. Jahn & B. Dunne

 

 

Through the REG’s available at the PEAR Lab Tim and I will utilize the mini Frog zamboni  table experiment to test various phenomena between random events and the operator. We will both test the influence of (an) operator(s) to keep the zamboni on the table and to push it off of the table. We will achieve this through running our own experiments and collecting our own empirical data to see what we find.  We hope to not only use ourselves in the lab but many of our friends and look to explore differences that may persist between males and females, groups of subjects that are single and mixed sex, as well as bonded couples.  We would also like to do this with the operator in the same room and as distant operators.  We anticipate finding correlations and some evidence of human influence on the frog zamboni driving off the side of the table and in remaining on the table.  We also expect to probably find higher levels of resonance amongst groups who are either bonded couples or close friends.  We have yet to set up and official meeting to flush out the finer details of our project with either Professor Jahn or Dunne, but hope to soon.  I did briefly speak with Brenda about the feasibility of such an area to explore and she enthusiastically thought this was a great experiment to try.

 
Gender Differences in Human Machine Interaction found in Princeton's
Pear Lab
Kelly Schaeffer
Professor Cooper/ Professor Jahn
 
I was very intrigued by my visit to the Pear Lab last week and by the
supposed findings that exist in gender differences in human-machine
interaction success. As a psychology major, I have found that gender
differences is a closely regarded topic in the psychology field and has
proved to be a very interesting one. I would like to review the findings
the pear lab has produced in their work on gender differences. I have
emailed Brenda Dunne to discuss the possibility of the project and she
has allowed me to stop by the lab to obtain the reports they have
created. I will review them and formulate a paper that will discuss the
findings and how they relate to other findings. Perhaps I can recruit
some friends to run through a few tests and then see if the results we
get match the ones already found by the lab. That would be phenomenal if
I could pull that off with the friends I have. The possibility of a
human influencing a machine amazes me and I wonder if either gender
truly has more of an influence. I am aware of the general gender
difference findings that Professor Cooper has lectured on, but I would
like to see the exact differences the lab has found after viewing the
experimental machines they have utilized. I would assume that males
would have more of an influence on machines as they seem to have a
greater affliction for them while females tend to be more intimidated by
machines. I am anxious to read of the exact results from the pear lab.
 
I apologize for the lateness. I was awaiting the response from Brenda
Dunne to okay the project. Thank you.
 
Kelly Schaeffer
*************************************8
 

Exploring Tic-Tac-Toe Learning with a Neural Network

Garo Hovnanian

Chris Asplund

Advisor, Reader: Jack Gelfand

 

Abstract

This project is designed to teach a neural network how to play Tic-Tac-Toe, and analyze the network’s representation of the game space to determine what, exactly, the network learned.  We will train our multi-layered network on many different Tic-Tac-Toe games, generated at both the novice and expert level.  Throughout training, the network will tune its sparse hidden layer representations to represent conjunctions of features on the input space that are the most relevant to playing the game.  This will be achieved with a mix of both Hebbian learning and top-down error-driven learning, the latter of which is based on the success or failure of a particular move.  Training will be considered complete when the network performs at an acceptable level of success on testing against both novice and expert players.  After training, we will analyze the receiving weights of the hidden layer from the input layer.  Presumably, these weights will represent which features of the input layer are important, as these features have become emphasized over the training epochs and other, irrelevant features have become deemphasized.  Thus, the shapes represented in the hidden layer will represent what this particular network learned using this particular training set.  We will then randomize the order of the training set and run the experiment again.  After a few such runs, we will extract the most relevant features of the input space that the hidden layer learned in all, or most, trials, and consider these features to be most representative of the shapes and features on the game board that are most relevant to being able to succeed in Tic-Tac-Toe.  The network starts out, through its randomly assigned sparse representations, recognizing the entire concept language of the game, but by the end, only represents the most important elements of the concept language.  We will compare the concept formations grown by the network to the concept formations that the Hoyle algorithm hard-coded into its network.  Thus, this experiment can either buttress or weaken the arguments made by the Epstein, Gelfand, and Lock paper as to what shapes are the most important to recognize in order to win in Tic-Tac-Toe.  In our presentation, we will have the network play against members of the class to see how well it fares.

 

 

J Bowen &  L Szolomayer

 

Studying Spatial Reasoning Using Matchstick Games

–or–

And Your Mother Said Never To Play With Matches

 

Advisor: Prof. J.J. Gelfand

 

Objective/Purpose:

Our project will study simple spatial relations using matchstick games.  Matchstick games are very simple, purely spatial games.  For example, starting with Figure 1, remove 2 “matches” (line segments) to create four equal squares in which all matches are used in the squares.  Figure 2 shows one possible solution to the problem

 

 

 

 

Figure 1

 

Figure 2

 
 

 


We hope to discover relationships between the starting configuration and the difficulty of the puzzle.  For example, configurations that start symmetrically and end asymmetric are more difficult then configurations that start asymmetric and end symmetric because our brain tends to favor symmetry.

 

Process:

In order to find the spatial relationships we will devise several starting configurations that lend themselves to the relationships we are studying.  For example, we will come up with several symmetric starting configurations and several asymmetric configurations.  Based on hypotheses we will develop we will “guess” which puzzles people will have a harder time solving and which ones will be easy.  We are planning on attending a talk on matchstick games on Wed. 4/23/03 to help us formulate our hypotheses as well as learn more about the different matchstick games.  After we have our test games we will have random test subjects complete the puzzles and then answer some simple questions to try to understand how the subject viewed the puzzle and their thought process in solving it. 

 

Anticipated Findings:

We hope to find, after gathering data from many test subjects, that there are certain spatial relations that change the difficulty of a visual problem.  We have not yet devised the specific hypotheses about the spatial relations because we are planning on attending a talk about the games on Wed. 4/23/03.  After the talk and extensively playing the game we will be able to formulate our relationship between game difficulty and spatial relations.

 

 
 

Nada Siddiqui

Final project description

 

Title: Bayesian Networks in Medical Diagnosis- How do machines and humans use probability theory to compute probabilities of having certain diseases?

 

Faculty member advisor: Professor Johnson-Laird

 

My project aims to construct a simple computer model of how computers use Bayesian network theory to compute the probability of having a specific disease, given certain inputs (symptoms, medical history etc). I will also conduct a psychological study of how humans use the same information at hand to calculate their own subjective probabilities and then compare both the results and the methods that humans and computers use.

 

I plan to implement a computer program, probably in C, that will ask a series of questions regarding symptoms and medical history of a hypothetical patient, take this information as inputs, and output the probabilities that the patient has a certain disease, with the program following a Bayesian network to compute the outputs. In order to limit the possibility of an NP hard problem, I will work with several inputs but limit the program to simple outputs (i.e. does have a disease or doesn’t, has type A skin disease or type B). After implementing the program, I will design and carry out a study, with Professor Johnson-Laird’s help, that will present information on conditional and prior probabilities of having certain diseases, given certain symptoms, and ask the person to come up with rough estimates of the actual probability of having a disease, given certain symptoms and histories. I will then compare the output of my program and the estimates to see whether people tend to over/underestimate this probability and try to come up with reasons as to why they do so.

 

 

Tyler Mincey

Karim Branford

PSY 322

21APR03

Human-Machine Interactions

Advisor:

 

 

Modern Applications of Wearable Computers to Provide Context-based Information

 

 

            Our project will investigate possible applications of wearable computers in modern society. Powerful computing power can currently be packaged in small portable units allowing innovative design opportunities for interactive systems that could potentially change the way information is remotely accessed. Our project will focus on wearable computers which can provide automated or semi-automated context-based information to the wearer.  A wearable computer can be outfitted with various sensors (GPS, cameras, microphones, etc.) which allow the machine to observe the wearer’s whereabouts and surroundings. Based on sensor readings the computer can provide environment specific information that it perceives to be needed by the wearer. We will research current methods of intelligent sensing (face/voice recognition, image/audio processing) as well as portable input/output devices (chord keyboards, small monitors, eye glasses adaptors) and propose system designs for the integration of these devices. Designs will focus on the ergonomics of the wearer/computer interface, power limitations, and the ability to autonomously deliver needed information. The social impact and information security of the suggested wearable computing applications will also be discussed.

 

Christina Mesner, 
Prof. Jahn
 
For my term paper project, I would like to study the effects of attention to the task on human-machine interactions. I am planning to use one of the Random Event Generators at the PEAR lab to find out if the human-machine interactions that have been observed at the PEAR lab are in any way correlated to how much attention the operator pays to the experiment. I will complete two sets of trials with the REG, one in which the operator will be concentrating on the experiment, and one in which the operator will still record the intention before the start of every run, but then focus on some other activity while the REG is running. I will then analyze the data to find out if there is any anomalous effect at all, and how the attention factor changed the outcome.