Abstract: An automatic agorithm for finding racing lines via computerized minimization of a measure of the curvature of a racing line is derived. Maximum sustainable speed of a car on a track is shown to be inversely proportional to the curvature of the line it is attempting to follow. Low curvature allows for higher speed given that a car has some maximum lateral traction when cornering. The racing line can also be constrained, or “pinned,” at arbitrary points on the track. Pinning may be randomly, deterministically, or manually and allows, for example, a line designer to pin the line at any chosen points on the track, such that when the automatic algorithm is run, it will produce the smoothest line that still passes through all the specified pins.

Abstract: A skill scoring frameworks allows for handicapping an individual game player in a gaming environment in preparation of matching the game player with other game players, whether for building teams or assigning competitors, or both. By introducing handicapping into the skill scoring framework, a highly skilled player may select one or more game characteristics (e.g., a less than optimal racing vehicle, reduced character capabilities, etc.) and therefore be assigned a handicap that allows the player to be matched with lower skilled players for competitive game play. Handicaps may apply positively or negatively a player's skill score during the matching stage. Handicaps may also be updated based on the game outcomes of the game play in which they were applied.

Abstract: A real-time stereo video signal of a captured scene with a physical foreground object and a physical background is received. In real-time, a foreground/background separation algorithm is used on the real-time stereo video signal to identify pixels from the stereo video signal that represent the physical foreground object. A video sequence is produced by rendering a 3d virtual reality based on the identified pixels of the physical foreground object.

Abstract: Skill scores represent a ranking or other indication of the skill of the player based on the outcome of the game in a gaming environment. Skills scores can be used in matching compatible players on the same team and matching opposing players or teams to obtain an evenly-matched competition. An initial skill score of a player in a new gaming environment may be based in whole or in part on the skill score of that player in another game environment. The influence that the skill scores for these other game environments may have in the skill score seeding for the new game environment may be weighted based on a defined compatibility factor with the new game environment. The compatibility factor can be determined based on a game-to-game basis, compatible categories or features, game developer defined parameters, or any combination of considerations.

Abstract: An adaptive pattern classifier makes use of training patterns and a known non-linear invariance transformation to generate a classifier representation based on an infinite set of virtual training samples on a training trajectory. Given the non-linear invariance transformation, optimization can be formulated as a semidefinite program (SDP), which is given by a linear objective function that is minimized subject to a linear matrix inequality (LMI). In this manner, a small training set may be virtually supplemented using the non-linear invariance transformation to learn an effective classifier that satisfactorily recognizes patterns, even in the presence of known transformations that do not change the class of the pattern.

Abstract: A modified large margin perceptron learning algorithm (LMPLA) uses asymmetric margin variables for relevant training documents (i.e., referred to as “positive examples”) and non-relevant training documents (i.e., referred to as “negative examples”) to accommodate biased training sets. In addition, positive examples are initialized to force at least one update to the initial weighting vector. A noise parameter is also introduced to force convergence of the algorithm.

Abstract: Scoring of each player may be based on the outcomes of a game between players who compete against each other in one or more teams of one or more players. The scoring may also consider partial play where one or more players of a game only play for a portion of the full time of the game. Additionally or alternatively, the scoring may consider partial rankings of teams where the relative ranking of one team to another may be unknown. Each player's score may be represented as a distribution over potential scores which may indicate a confidence level in the distribution representing the player's score. The scoring may be used to track a player's progress and/or standing within the gaming environment, used in a leaderboard indication of rank, and/or may be used to match players with each other in a future game.

Abstract: Racing-based computer games typically include a mode in which one or more human players can compete against one or more computer-controlled opponents. For example, a human player may drive a virtual race car against a computer-controlled virtual race car purported to be driven by Mario Andretti or some other race car driver. Such computer controlled opponents may be enhanced by including a sampling of actual game behavior of a human subject into the opponent's artificial intelligence control system. Such a sampling can allow the game system to personalize the behavior of the computer control opponent to emulate the human subject.

Abstract: Based on the time series data from multiple components, the systems administrator or other managing entity may desire to find the temporal dependencies between the different time series data over time. For example, based on actions indicated in time series data from two or more servers in a server network, a dependency structure may be determined which indicates a parent/child or dependent relationship between the two or more servers. In some cases, it may also be beneficial to predict the state of a child component, and/or predict the average time to a state change or event of a child component based on the parent time series data. These determinations and predications may reflect the logical connections between actions of components. The relationships and/or predictions may be expressed graphically and/or in terms of a probability distribution.

Abstract: Players in a gaming environment, particularly, electronic on-line gaming environments, may be scored relative to each other or to a predetermined scoring system. The scoring of each player may be based on the outcomes of games between players who compete against each other in one or more teams of one or more players. Each player's score may be represented as a distribution over potential scores which may indicate a confidence level in the distribution representing the player's score. The score distribution for each player may be modeled with a Gaussian distribution and may be determined through a Bayesian inference algorithm. The scoring may be used to track a player's progress and/or standing within the gaming environment, used in a leaderboard indication of rank, and/or may be used to match players with each other in a future game.

Abstract: Adaptive agents are driven by rewards they receive based on the outcome of their behavior during actual game play. Accordingly, the adaptive agents are able to learn from experience within the gaming environment. Reward-driven adaptive agents can be trained at either or both of game-time or development time. Computer-controlled agents receive rewards (either positive or negative) at individual action intervals based on the effectiveness of the agents' actions (e.g., compliance with defined goals). The adaptive computer-controlled agent is motivated to perform actions that maximize its positive rewards and minimize is negative rewards.

Abstract: An adaptive pattern classifier makes use of training patterns and a known non-linear invariance transformation to generate a classifier representation based on an infinite set of virtual training samples on a training trajectory. Given the non-linear invariance transformation, optimization can be formulated as a semidefinite program (SDP), which is given by a linear objective function that is minimized subject to a linear matrix inequality (LMI). In this manner, a small training set may be virtually supplemented using the non-linear invariance transformation to learn an effective classifier that satisfactorily recognizes patterns, even in the presence of known transformations that do not change the class of the pattern.

Abstract: Improved human-like realism of computer opponents in racing or motion-related games is provided by using a mixture model to determine a dynamically prescribed racing line that the AI driver is to follow for a given segment of the race track. This dynamically prescribed racing line may vary from segment to segment and lap to lap, roughly following an ideal line with some variation. As such, the AI driver does not appear to statically follow the ideal line perfectly throughout the race. Instead, within each segment of the course, the AI driver's path may smoothly follow a probabilistically-determined racing line defined relative to at least one prescribed racing line.

Abstract: Racing-based computer games typically include a mode in which one or more human players can compete against one or more computer-controlled opponents. For example, a human player may drive a virtual race car against a computer-controlled virtual race car purported to be driven by Mario Andretti or some other race car driver. Such computer controlled opponents may be enhanced by including a sampling of actual game behavior of a human subject into the opponent's artificial intelligence control system. Such a sampling can allow the game system to personalize the behavior of the computer control opponent to emulate the human subject.

Abstract: In a virtual reality environment, the behavior of the computer-controlled virtual vehicle may be made more human-like by increasing the AI driver's reaction time to environmental stimuli, such as physical stimuli (e.g., detecting a loss of tire traction, audio warning signals, smoke, virtual fatigue, weather changes, etc.) or “visual” stimuli (e.g., virtual visual detection by the computer driver of a turn or obstacle in its path, ambient lighting differences, etc.). Reaction time may be increased by introducing a delay in receipt of stimuli by the artificial intelligence motion control system, by introducing a delay in receipt of control signals by the physics engine, or by modifying the control signal to degrade their accuracy in approximating a prescribed racing line.

Abstract: Improved human-like realism of computer opponents in racing or motion-related games is provided. The computer driver may be “distracted” by various characteristics, such as nervousness caused by another competitor closing the gap behind a computer driver. The distraction effects may be reflected in the alteration of stimuli to represent the computer driver “missing” stimuli, as though the AI competitor has taken its virtual eyes away from the course in front of its vehicle for an extended time period in order to watch the racing vehicle behind it. In addition, some distractions may be caused by different directional stimuli. When it is determined that an AI driver has glanced into the rear view mirror, visual stimuli from in front of the vehicle may be skipped because a human driver would not be able to simultaneously process visual stimuli from both the front of the vehicle and the rear view mirror.

Abstract: A modified large margin perceptron learning algorithm (LMPLA) uses asymmetric margin variables for relevant training documents (i.e., referred to as “positive examples”) and non-relevant training documents (i.e., referred to as “negative examples”) to accommodate biased training sets. In addition, positive examples are initialized to force at least one update to the initial weighting vector. A noise parameter is also introduced to force convergence of the algorithm.