Wednesday, November 30, 2011

ICIDS Day 1 - Paper Session #3

Paper Session #3: Virtual Characters and Agents

Talk 1: A Knowledge-Based Framework for the Collaborative Improvisation of Scene Introductions

How do improvers do their stuff? i.e. reach shared understanding, reason about the scene, co-create a narrative?
Built a game called Party Quirks
Three Line Scene? Tiny West?

Talk 2: A New Approach to Social Behaviour Simulation: the Mask Model

Social performance – different roles in different contexts
Each mask contains a level of influence from 0 to 1, for example, religion as a mask – a low value would describe a moderate believer, while a high value would correspond to a fanatic.
If a character has a certain role, when a pre-condition is triggered, an action on an entity is suggested with a certain strength, for example, if I come here to present, I could get up here and be silent, but the strength of the pressure to present is quite high.
It is also important to note that the character will usually have access to the rules denoting other character roles, even without possessing those masks themselves, e.g. we all know we should not take other people’s computers, even though few of us may actually be thieves.
There are three layers to each character – self-perception layer (can model self-deception) is always active, social layer (including religion, politics, culture, etc) which includes beliefs and their respect for social constraints, and the interpersonal layer (persistent, or only active in the presence of a certain character, e.g. mother, girlfriend or boss).
Need to manage the conflicts between the masks – e.g. Romeo’s different masks say he should love Juliet, but his other mask that he shows toward his parents says no. How do we decide how he will act in a given situation?
What is the perceived possibility of him being in a given social situation where this conflict will arise? What are the pros and cons? What is the interpreted morality of each action, and which one better suits with the needs of the character?
Ego-esteem – desire to do the thing, moral soundness of the thing, possibility of getting caught
When two opposing factors result in one of the factors being subverted because the other is far stronger, this becomes a dramatic turn in the plot.
 It’s more efficient to create stereotypical masks, then assign several of those to each character, rather than custom-designing each mask for each character.
This could also be used to create personalized NPCs, based on interpreting the moral actions of the PC within the culture of the simulated world.
Murray – “This is akin to modeling all the muscle systems of a dog so you can get an accurate representation of a dog’s choice of actions, but the cartoon dog, that behaves like a dog, is far more convincing. Where is the value for the interactor in including all of these extra systems?”

Talk 3: Perceived or Not Perceived: Film Character Models for Expressive Natural Language Generation

Dialogue generation – I want to flatter you, I’m a friendly person, but I’m feeling hesitant. Can you generate dialogue that reflects all of that?
Deconstructing elements of language to provide subtext – long or short sentences meaning terseness or pontification, use of larger words indicating education, etc.
More parameters = better dialogue?
… Why try to generate natural dialogue by creating caricatures of particular stereotypes and applying them to generic dialogue? Someone has to write that generic dialogue in the first place.

Talk 4: Representing Dramatic Features of Stories through an Ontological Model

Extracting generic information from video clips – i.e. actions, characters, motivations, content
Automatic segmentation of narrative units? Automatic storyboarding and previs?
DMO – Dramatic Media Object?
Polti’s situations?

ICIDS Day 1 - Paper Session #2

Paper Session #2: New Authoring Modes

Talk 1: A Method for Transferring Probabalistic User Models Between Environments

Choice is related to preference; mathematical psychology tells us preference is related to utility – changes in choice behaviors are reflected in changes of utility and vice versa
The Strict Utility model – determining preference given utility
Reciprocity – if I do something nice for you, even unsolicited, you’re more likely to do something nice for me. Can this be used in IF? Is it a case of ‘liking’?
The Fechnerian Utility model – if you ask someone to commit to doing something, then come back and ask them to do it again, they’re more likely to do it. The comes from Greenpeace – doing a mailout, they would get 18% donations. If they sent a mail 3 weeks earlier that asked if they were committed to the environment, then did their mailout, they received 35% donations. – CIIGAR Lab

Talk 2: Being in the Story: Readerly Pleasure, Acting Theory and Performing a Role

Digital stories are not created by the actor, they are enacted – this means there is no conflict between interaction and narrative.
Readerly pleasures – loss of agency vs. authorial (interactive) pleasure – unrestricted agency
IF research focuses on authorial pleasure because the researchers are almost always writers/designers.
More research needs to be done into method acting as being similar to what a player undergoes when enacting their character.
Method acting is not having had the experience, but in taking pieces of one’s experiences and transmuting them into an approximation of an experience. This makes the process imaginative, rather than just mnemonic.
… This was a PhD confirmation presentation, not a paper.

Talk 3: Supporting Rereadabillity Through Narrative Play

How to encourage repeated, satisfying experiences of interactive stories?
Reread for variety, for closure, for achievements?
Challenged-based games – replay to do better (get a higher score)
Card game – Once Upon a Time – be the first one to finish the story
But how to keep narrative from becoming subordinate to gameplay? Create narrative moves and gameplay moves, and make it impossible to win through only one kind of move or another.
Introduce narrative and gameplay elements, then remove those once they are carefully planned and watch the participants react – they want to do better next time
Jones in the Fast Lane…
Challenge connected to the story – motivated to do better next time
The cards will be available for download eventually

Talk 4: Extensible Tools for Practical Experiments in Interactive Digital Narrative

ASAPS engine
Tech demo? Ren’Py engine?

ICIDS Day 1 - Paper Session #1

Paper Session #1: Interactive Storytelling Theory

Talk 1: Research in Interactive Drama Environments, Role-Play and Storytelling

Why so few complete narrative systems? Fa├žade and what else?
“Story management as defence of authorial vision.” As opposed to:
“Story management shapes dynamic generation of engaging experience.”
Storyfication of the experience is stored in the participant’s autobiographical memory.
Creation of dramatic structure (ensured, but brittle), versus creation of characters (no dramatic structure ensured, but very flexible).

Talk 2: Why Paris Needs Hector and Lancelot Needs Mordred (Janet Murray! – Georgia Tech)

New book – Inventing the Medium
I consider myself a recovering AI researcher, though I’m having serious relapses.”
Reliving Last Night – 2001 – Sarah Cooper
Vladimir Propp – Morphology of the Folktale (1928)
The tale of two boyfriends – ties of duty vs. the forbidden sexy guy
The doubling up of Hector/Achilles over Paris/Meneleus greatly heightens the dramatic tension
The double betrayal of Guinevere and Lancelot to Arthur creates tension, as does their initial refusal of each other – Mordred is Lancelot’s dark side. He just wants Guinevere, and to take over the kingdom, while Lancelot is virtuous.
The Miller’s Tale? The foils, in effect, heighten the characteristics of the original forbidden sexy guy. E.g. Wickham vs. Darcy.
Abstract roles and functions to create a series of classes that can be applied to a moral sliding scale, so that characters may foil each other at multiple points while creating meaningful character interactions, e.g. the number of different proposals Elizabeth receives in Pride & Prejudice.
Cultural roles are systems of abstraction – you do not need to reproduce the social world, you want to reproduce story atoms, not social knowledge which already exists within these story atoms.
Agency does not come from freedom to do anything, agency comes from scripting interactions that are meaningful and suitably rewarding.
Telling more complex stories expands our humanity.

Talk 3: Agent-oriented Methodology for Interactive Storytelling

Fuzzy cognitive maps for exploring causal-related concepts (inference engine).
Done in Oblivion AI – scaled back because it made the world too unstable.
Also more recently done in The Snowfield by MIT Gambit Games Labs.

Talk 4: Back-Leading through Character Status in Interactive Storytelling

User agency vs. authorial control – i.e. will a player as Anna Karenina be willing to kill their character to make for a better story?
Why do we expect novice storytellers (i.e. players) to create a good story when provided with no clues from the designers as to how to proceed? Especially when it takes writers years to acquire the skill that allow them to tell a good story in the first place?
Improvisational theatre – e.g. Tiny & Tony’s Wedding, Turtle Talk with Crush – designed to act with novice audiences who have little experience in creating stories.  You can sometimes provide input, but it’s still designed to be observed.
Inter-actors vs. spect-actors – back-leading is leading while appearing to follow
Interactive theatre – designed to be experienced, not observed, e.g. The Second City
Status shifts create drama – changes in dominance or submission, one-upmanship, etc – use status to back-lead the audience without explicitly telling them how to react
Force Dynamics in Language and Cognition
Could very easily be used in one-on-one confrontations between player and character!
E.g. civilians in WoW – “Too good to wait in line like the rest of us, Mage?” prompts the answer of “Yes.”
(Performance) Keith Johnson, “People tend to minimize status gaps.”? Does this still apply in virtual environments?

Talk 5: Rereading in Interactive Stories: Constraints on Agency and Procedural Variation

IF tries to maintain agency, immersion and transformation, which leads to the need for variety across instances that, nevertheless, maintain coherence within themselves.
Reframing: The Sixth Sense, BioShock – the twist that changes everything and makes people want to re-experience
What additional constraints does this impose on the variations you can create when the reader has this new knowledge and they’re re-watching/re-playing?
You need to maintain coherence not just within, but across all sessions.
Events relevant to the reframing must remain the same, e.g. Bruce Willis can never be not-dead.
Some player/reader actions may be limited by revelations that have not yet happened – why can the character not do certain things? How do you communicate this without ‘giving up the ghost’?
There are also constraints upon what must be omitted, or what can be omitted.
The reader will be looking for different things, e.g. if the anniversary dinner scene wasn’t in The Sixth Sense the second time you watched it, you would feel disappointed, because you want to find out how you were fooled.
Constraints on the ordering – the reframing must occur in the first reading, otherwise there is no impetus to re-experience. Fight Club?
There needs to be sufficient discourse time between the reframed event and the reframing – if The Sixth Sense was only 10 minutes long, you wouldn’t need to re-experience it, because you could remember.
Versus constant reframing, i.e. Memento?
Reframing in Heavy Rain? How did I miss seeing who the killer was? Coherence lacking – when did he go into the back room? Why is the moment prolonged? Is it there? It seems natural the first time, but the second time suspicious. Does he leave twice?
Unreliable narrators? How do these work in IF? The Jade Smiley story – based on King Lear

ICIDS Day 1 - Keynote: Keith Oatley

Keith Oatley – Stories as Simulation in Print Fiction, Movies, and Interactive Media

University of Toronto – Physiological psychology, HCI, novelist
Fiction: something that’s been carefully constructed to create a certain psychological outcome
Writers of fiction were continuing the play of childhood – all mammals play, but not lizards or birds (how sad!). Brian Boyd, in 2009, said play is the origin of stories.
Art in general is the more or less conscious externalization of metaphors – this thing is something that it isn’t. A banana is also a phone to a child in play.
Paint factory 100,00 years ago; beads from 82,000; flute from 43,000, ritual burial from 40,000; cave paintings from 31,000.
The power of fiction is that metaphor can apply to oneself, as well as the world.
Donald Winnicot – Playing and Reality
Throughout childhood, one learns to define oneself by conceptualizing the other (caregiver) as a comparison to what one is not. This is mentalization. We understand others as we understand ourselves.
Beatrice Beebe – mother-child interactions on a second-by-second basis – the prototype of play
Interactive fiction beings in infancy – it is indivisible from play.
“Fictional stories are simulations that run on minds.” (Oatley, 1999 – Why Fiction May Be Twice as True as Fact)
“Such Stuff as Dreams” – The Psychology of Fiction – Keith Oatley
Fiction has binocular vision – we see the surface level, but we also understand some of what happens beneath the surface, and all of this is based on what we see in others and in ourselves, within the interactive frameworks we’ve come to recognize throughout our lives.
Flint & bags – 2,500,000 years ago; conversation – 250,000 years ago; writing – 5,000 years ago; printed books – 500 years ago; computers – 50 years ago; iPhone – 5 years ago.
There is a scale of involvement from a roller coaster ride (passive) to a novel (active emotional involvement)
Play: interact with another; games: carry out (competitive) action; oral storytelling: engage with storyteller; theatre: engage with drama model; print fiction: identify, sympathise; film: identify with desire; video games: carry out action, visit scenes; digital narrative: choice of action
Reflection, skills, listening, audience/discussion, reflection/discussion, listening/audience, skills, reflection
“Stories are trajectories of desire.”
Patrick Hogan (2003) – three kinds of stories are human universals – love story, heroic story (anger & conflict), suffering & sacrifice for the community
Fiction isn’t description – the writer’s job is to offer cues/instructions – a metaphor that evokes an idea. In turn, the reader’s part is to mentally create the imagined world – to start up your simulation and keep it running. They need to bring alive a simulated world and take a personal part in it, otherwise it’s all just marks on paper.
Each technology has three parts – the external object, the skills required for use and the cultural conventions surrounding that use.
You only really need three objects to create a vivid scene – people took cues of 6 phrases that described a room, but brain activity decreased after the first three, to the point where it is possible to infer that the room is as vividly imagined as necessary at that point.
Demis Hassabis – rising star?
The brain takes clues and applies to them previous understanding – it follows models and then reapplies them to the current situation. (time-saving processes - semiotics)
Fiction piggy-backs on this ability of the mind to recognize cues and create something understandable.
Kuleshov effect – if you want to scare the audience, don’t have the actor look scared – have him look passive, then cut to the view of the object – this creates metonymy – feeling something that isn’t actually there.
Current rate of changing shots in movies is about one shot per 4 seconds
The real interface is not between us and the visual world, but between us and each other, primarily language.
Ilya Sutskever – neural networks for predicting upcoming text
“What relation would you like between understanding how to make things happen (computer programming) and understanding how to enable things that happen (writing simulations that run on minds)?
What you want to be able to do is enabling people to feel what they feel, to think what they think, not telling them when and what to feel and think and how.
Psychology of Fiction –
If using a flight simulator can make you a better pilot, reading fiction could make you better at socializing (simulating environments). Do people who read fiction have improved theory of mind?
The more fiction you read, the better you are at recognizing cues given by people’s eyes (Simon Baron Cohen’s test). The more non-fiction you read, you’re not necessarily better, but it’s an expertise factor. Reading about genetics will make you better at identifying genetics, so reading fiction about relationships will make you better at identifying relationships.
Maja Djikic  - reading Chekov’s story The Lady with the Little Dog and a non-fictional version
After reading, each person changed a little. More importantly, each person felt and changed in their own way, not in a uniform emotion or sense, but in a way that meshed with their personality.
“Art enables us to change, not in a way desired by others, but in our own way.” Chekov believed ‘stating a problem correctly’ was better for an artist than ‘solving a problem’.
Fiction is not making things happen, it is enabling things to happen. [Paraphrased] Fiction gives us the ability to read within ourselves.
Avril Thorne – narrative psychologist
“As a psychologist, I can tell you what to think in a given situations. As a novelist, I don’t want to do that – I want to say, “Here’s this situation, what do you think?”
Stories do not end – culminations occur to episodes of interaction.