Homework 3

            This could have been a very interesting paper had the author written it very differently. I did not like the style he wrote the paper and the way he explained his points. This paper was about minds brains and programs. What the paper was about was the author attempting to differentiate between strong and weak AI and explaining what he thought would be strong AI. Weak AI is mainly just a tool but according to the author strong AI is a mind and can learn and think for itself. He starts the paper by explaining in detail what he thinks the definition of strong AI. He mentions a something about humans understanding stories; he tells a story about a man in a restaurant and explains that even if something is not said in a story a human would be able to infer that something happened and that strong AI would be able to do the same thing.

            After that he explains a way that would test the theory, the way to test it is called the Chinese room. What that is is lock a person into a room who does not understand Chinese and you give them papers with Chinese writing on them. Then after this first batch of papers you are given a second batch but with this one you are given a sheet of rules on how to match up the symbols in English. To someone sitting outside the room it looks as if you understand Chinese and are able to read the stories that are presented. The point of this is to show the basic way programs would work in “understanding” things. In this situation you are the program, matching symbols with a list of rules and the user is the person giving you those symbols and rules.

            After this example and explanation the author decides to poll different people in the AI field as to what strong AI is exactly. Then he basically shoots everyone who replied down saying that he was right and there was no way that they were right. At one point he says “This objection really is only worth a short reply” and gives a very short paragraph of how he was right and that the reply was just pointless.

            The objection he said that to, in my opinion was actually one of the better ones, what it said was "How do you know that other people understand Chinese or anything else? Only by their behavior, now the computer can pass the behavioral tests as well as they can (in principle), so if you are going to attribute cognition to other people you must in principle also attribute it to computers.” This statement actually makes sense to me, what I get from it is that you don’t see the people understanding what you are saying or in what language you know by their behavior that they do in fact understand you and if a computer can do that then that means that the computer is understanding you.

            Overall I didn’t really like this paper, it was not well written and the author seemed like he didn’t care what the other people he polled said they were wrong and he was right. So that kind of took away from the paper and made it not a very good read for me.

Paper Reading #6: Implanted user interfaces

Introduction
Title: Implanted user interfaces
Author Bios:

• Christian Holz-Autodesk Research, Toronto, Ontario, Canada & Hasso Plattner Institute, Potsdam, Germany
• Tovi Grossman-Autodesk Research, Toronto, Ontario, Canada
• George Fitzmaurice-Autodesk Research, Toronto, Ontario, Canada
• Anne Agur-University of Toronto, Toronto, Ontario, Canada

Summary

In this paper they discussed implanting user interfaces underneath human skin. They went over 4 of the main obstacles that are coupled with implanting devices under the skin, they were input, output, communication, and power supply. They came up with several solutions for each category and tested them on a cadaver. For input they tested a button, pressure sensor and a tap sensor. For output they tested LED, a vibration motor, and a speaker. They only tested Bluetooth for communication even though they though WiFi would work also. For a power source they had an inductive charger recharging a battery. This was not the actual study, all of this testing was to make a prototype that could be tested on users. The prototype was placed under artificial skin and users tested it out while doing other activities.

Related Work

The papers they referenced in the paper were more or less explanations of how this is a novel idea and things that explained an idea that they had in the paper so there are not many papers on related work to implanted user interfaces. There are papers on user interfaces but none on implanted user interfaces even in the paper they say "Despite these potential benefits there had been little or no investigation of implanted user interfaces from an HCI perspective."

• A miniaturized tunable microstrip antenna for wireless communications with implanted medical devices
• Wireless monitoring of electrode-tissues interfaces for long term characterization
• Towards an activity-aware wearable computing platform based on an egocentric interaction model

Evaluation

The way they evaluated the actual study was they had 4 users wear the artificial skin with the prototype underneath it and had them go out and public and do certain things. The prototype would interrupt them with a game that it had set up and they would try to get a high score. After about an hour they asked the user how easy it was to use the device and which input/output devices they liked the most. So the overall evaluation was qualitative, they asked the users what they thought of the device. What they found was that people disliked the pressure sensor and the LED was hard to see in the light.

Discussion

 I thought that the idea of implanted user interfaces was very interesting. I pictured people walking around texting on their arm, or checking email in their hand. They explained in the paper that this was a novel idea, people had looked into worn interfaces and things like that but never implanted interfaces that provided some sort of feedback.  

Paper Reading #4: Profanity use in online communities

Introduction
Title: Profanity use in online communities
Author Bios:

• Sara Sood-Pomona College, Claremont, CA, USA
• Judd Antin-Yahoo! Research, Santa Clara, California, United States
• Elizabeth Churchill-Yahoo! Research, Santa Clara, California, United States

Related Work

• Designing for improved social responsibility, user participation and content in on-line communities
• Automatic satire detection: are you having a laugh?
• A life-cycle perspective on online community success
• A multilevel analysis of sociability, usability, and community dynamics in an online health community
• User loyalty and online communities: why members of online communities are not faithful

From what I read this isn't really a novel idea, people have known that online communities have bad systems that regulate profanity and things like that.

Summary

In this study the authors took comments from a now non-existent website (Yahoo! Buzz) and had people comment on whether or not there was profanity in the comment or if it was an insult and if it was an insult who was it towards. They took about 6500 comments from that website and had about 200 people go through a few at a time and answer those 3 things, they had a thing called gold comments that had correct labels on them so they would know if the people were confused or not doing the study right. If they got too many of the gold comments wrong they would throw out their data. After they got the results from that they took lists from a couple of sites of profane language and setup a detection system so they could test how accurate the systems people employ now are. 

Evaluation  

To evaluate they took the answers from the people actually going through each comment and compared that data to the system that they set up to go through the comments. I would say that this is a mixture of qualitative and quantitative methods since they got data by asking questions and compared the data using numbers. They found that the systems that websites use today aren't really as good as they need to be since they miss things that are spelled wrong or when people use symbols to finish spelling the words. They also explained why the current systems are not adequate. 

Discussion

This was a little interesting but not much since it is obvious everywhere on the internet that people easily get by systems that ate supposed to find profanity and get rid of it. It was interesting to see why they miss those things though, I didn't know that they use set list that rarely get updated with the new slang that comes out all the time which makes sense since you always see people finding new ways to insult other people on forums and things like that. It also didn't seem very novel as they mentioned people who had already gone over some of this stuff.

Paper Reading #5: Looking glass: a field study on noticing interactivity of a shop window

Introduction
Title: Looking glass: a field study on noticing interactivity of a shop window
Author Bios:

• Jörg Müller-TU Berlin, Berlin, Germany
• Robert Walter-TU Berlin, Berlin, Germany
• Gilles Bailly-TU Berlin, Berlin, Germany
• Michael Nischt-TU Berlin, Berlin, Germany
• Florian Alt-University of Stuttgart, Stuttgart, Germany

Related Work

• Through the looking glass: you can play against your own reflection
• The Looking Glass IDE for learning computer programming through storytelling and history exploration: conference workshop
• Alice on both sides of the looking glass: Performance, installations, and the real/virtual continuityThrough the looking glass: the use of lenses as an interface tool for Augmented Reality interfaces
• Through the looking glass: the use of lenses as an interface tool for Augmented Reality interfaces
• Through the looking glass of immaterial labor
• Through the looking glass: game worlds as representations and views from elsewhere
• Chained displays: configurations of public displays can be used to influence actor-, audience-, and passer-by behavior
• How to evaluate public displays
• Interactive television: new genres, new format, new content
• Exploring factors that influence the combined use of mobile devices and public displays for pedestrian navigation

This work doesn't seem very novel as there are a lot of papers out on this subject, but they did talk about related work well in the paper.

Summary

This study was mainly about seeing what factors contributed to noticing the interactivity of something. They had 3 separate studies from which they collected data. They had a pre-study where they set up a prototype at a university to see how people would notice and what they would do. What they got from the pre-study was that people tend to interact in groups and they stopped to interact rather than doing it as they passed by. The next study was a lab study where they just tried to see how long it would take for someone to realize that a display was interactive. The final study was a field study where they set up a interactive display in a public area for 3 weeks and changing it a little each week to see how different user representations effected the time it took to realize the interactivity.

Evaluation

In the paper they said that they collected both qualitative and quantitative data. Qualitative from observations, interviews, and video recording. Quantitative from complete interaction logs and videos from a depth camera. What they found was that between a mirror image, a silhouette, and an avatar representation the mirror image was the thing that got the most attention the quickest. They also found that if one person starts interacting that more and more people will start to interact with the display since they see that it is interactive so easily. 

Discussion

This topic was very interesting, it didn't seem very novel as there findings seem to have been obvious. It takes awhile for people to notice something is interactive and if they see someone else interacting with it then of course they are going to go play with it too, it seems like human nature to me. The only novel part was that  instead of having something on the screen saying that it was interactive they tried to see how long it would take without that. Overall it was a very interesting study though, I would like to see what else they do with things like this.

 

Paper Reading #3: Protecting artificial team-mates: more seems like less

Introduction
Title: Protecting artificial team-mates: more seems like less
Author Bios: Tim Merritt National University of Singapore, Singapore, Singapore
Kevin McGee National University of Singapore, Singapore, Singapore

Summary

In this study they had participants play a video game with a AI and then had them play with a presumed human(PH) which was actually the same AI doing the same exact thing. The purpose was to see how the players played when they thought they were on a team with an AI versus when they thought they were on a team with another human. The point of the game was to have both players touch the gunner in the middle, they didn't have to touch at the same time but both had to touch to move on to the next level. The gunner scanned in a circle until it reached a player and then fired, the player could distract the gunner with the 'W' key or they could run into the field of vision of the gunner. What the study was on was how much the player hit the 'W' key to protect the AI or PH.

Related work

• A Failure of Imagination: How and Why People Respond Differently to Human and Computer Team-Mates.
• Proactive information exchanges based on the awareness of teammates' information needs
• Human-centered design in synthetic teammates for aviation: The challenge for artificial intelligence
• What we have here is a failure of companionship: communication in goal-oriented team-mate games
• Choosing human team-mates: perceived identity as a moderator of player preference and enjoyment
• Real-time team-mate AI in games: a definition, survey, & critique
• Using artificial team members for team training in virtual environments
• Can computers be teammates?
• The media equation: how people treat computers, television, and new media like real people and places
• Are computers scapegoats?: attributions of responsibility in human-computer interaction

This paper is novel as there is no other paper on this specific topic. They did reference papers correctly 

Evaluation

To evaluate the study they used several questions with a Likert-type scale so it was a quantitative and subjective approach to evaluate the results. They also used a qualitative question asking who the player protected more and why. What they found was that even though players protected the AI more they said in the questionnaire that they protected the human more. They then asked the Likert-type questions to observe stereotypes  and personal pressures. They also had players watch videos of 2 AI, 1 AI and 1 PH, and explain their behaviors in an open response so they could see how the players thought a human acted over how an AI acted.

Discussion

I thought this contribution was actually pretty interesting, since when starting the paper I thought that the players would naturally protect the human players more but was proven wrong when the opposite happened. Also they fact that the players thought they were protecting the humans more too even though they in fact protected the AI more. It is novel since no one has ever studied this field in this way.

Paper Reading #2: Touché: enhancing touch interaction on humans, screens, liquids, and everyday objects

Introduction
Title: Touché: enhancing touch interaction on humans, screens, liquids, and everyday objects
Author Bios: Munehiko Sato-Disney Research, Pittsburgh, Pittsburgh, Pennsylvania, USA & The University of Tokyo, Tokyo, Japan
Ivan Poupyrev-Disney Research, Pittsburgh, Pittsburgh, Pennsylvania, United States
Chris Harrison-Disney Research, Pittsburgh & Carnegie Mellon University, Pittsburgh, Pennsylvania, United States

Summary

What they did in this paper was they tried to find a different form of capacitive touch sensing, they called it Touché. Instead of using the normal capacitive touch sensing they used a novel form called Swept Frequency Capacitive Sensing (SFCS). This is different because rather than having a conductive object excited by a signal at a fixed frequency it uses a range of frequencies which would let them measure a lot more data points. Using this they testing how Touché could be used in the world today and how it could improve touch based systems. They even tested making liquid a touch surface, seeing how many fingers were in the water or seeing if only the surface was touched. They tested 5 domains with their new system: making everyday objects gesture sensitive, sensing human bi manual hand gestures, sensing human body configuration, enhancing traditional touch surfaces, and sensing interaction with unusual materials.

Related Work


Evaluation
They used a qualitative unbiased approach to the evaluation. They had the people train gestures and then proceeded to see if Touché could pick up that certain gesture with slight differences and measure the accuracy that way. They then took out the worst gesture to see how good the system could be at different things. They tested a door knob, body configuration, enhancing a touchscreen, on-body gesture sensing, and touching liquids.

Discussion
I thought their work was very interesting and extremely novel, I have never heard of any system able to see if you are touching a liquid or if your hand is submerged. I think the research presented in this paper will help innovate the touch sensing world and make way for new ways to tell a computer what to do and what gestures can be recognized.

Paper reading #1: Oh, dear Stacy! Social interaction, elaboration, and Learning with Teachable agents

Introduction
Title: Oh, dear Stacy! Social interaction, elaboration, and Learning with Teachable agents.
Author bios:
Amy Ogan- is a postdoctoral student at Carnegie Mellon University in the Human-Computer Interaction Institute. She works on virtual agents for learning with other students
Samantha Finklestein- is a doctoral student at Carnegie Mellon University in the Human-Computer Interaction Institute.
Elijah Mayfield- is a doctoral student at Carnegie Mellon University in the Human-Computer Interaction Institute.
Claudia D'Adamo-
Noboru Matsuada
Justine Cassell

Summary

What they did was basically observe how students from different grades interacted with a teachable agent that could supply a few different social responses to the child. They were trying to see what would have the most learning gain and how the way the children interacted with Stacy affected their learning gains. The hypothesis was: how do increased cognitive reflection moves, inside-system vs. outside-system language and increased social moves correlate with learning. The agent Stacy was to be taught linear equations from the child and the authors would observe how the child conversed with Stacy, if they called the agent "she" or "her" it was inside, if it was called "it" it was outside. They then evaluated the results by comparing how the student talked to Stacy and how much she learned or didn't learn and how much they learned.

Related Work

The work was novel and the way the authors talked about related work was appropriate and helped with the overview of what has already been looked at.

Designing learning by teaching agents: The Betty's Brain system

Virtual peers as partners in storytelling and literacy learning

A social-cognitive framework for pedagogical agents as learning companions

Measuring self-regulating learning skills through social interactions in a teachable agent environment

Modeling student behaviors in an open-ended learning environment.

A Science Learning Environment using a Computational Thinking Approach

Identifying Learning Behaviors by Contextualizing Differential Sequence Mining with Action Features and Performance Evolution.

Supporting Student Learning using Conversational Agents in a Teachable Agent Environment.

Relating Student Performance to Action Outcomes and Context in a Complex, Choice-Rich Learning Environment.

Identifying Students Characteristic Learning Behaviors in an Intelligent Tutoring System Fostering Self-Regulated Learning.

All of these papers do with learning from a teachable agent or interactive program, but none of them talk about what the authors of this paper talk about. Which is how social interaction can affect learning from a teachable agent

Evaluation
Their results were evaluated systemically, they took all of the data they collected from each child put it together and created a coding scheme that was applied to the things the children said aka "utterances" as they were teaching Stacy. The categories the utterances were put in were: a social utterance, a tutoring move, an alignment bases pronoun use, a cognitive assessment,a correctness evaluation and they also had a none category. Once they put all the utterances into the right category they looked for correlations between the categories and the learning gains. They also looked at shifts in behavior and how specific behaviors in the chilled affected upcoming alignment on a turn by turn basis. Since they took down the words the child said I think it was also evaluated quantitatively with a little bit of qualitative since they measured the learning gains with a pre and post test and found a normalized gain.

Discussion
I thought that the work and contribution were very interesting, I didn't think that teaching someone helped you learn but when I thought about it it made perfect sense. Also I didn't know that there were teachable agents that help with this. The way they evaluated the results seemed to be pretty appropriate to me since I couldn't think of another way to look at the data they acquired. The contribution was novel in my opinion the field has been researched but not in the way these authors did, the overall contribution should help with the future development of teachable agents to help kids learn in a different way.