US Patent 7549688 - Robotic hand with increased finger flexibility

http://www.freepatentsonline.com/7549688.html

Human fingers are require a delicate balance between flexibility and sensitivity to properly interact with the environment. This patent from Honda may provide their ASIMO robot with similar attributes. Claim 1 reads:

1. A robot hand having a finger comprising finger links and finger joints interconnecting the links, wherein at least one of the finger links comprises:

a base to which a force sensor is attached;

a first member connected to the base through the force sensor;

a second member adhered to a surface of the first member;

a third member covering at least the second member; and

a fourth member adhered to a part of a surface of the third member, wherein the fourth member is interconnected with the force sensor through a connector.

US Patent 7548894 - Neural network hardware

http://www.freepatentsonline.com/7548894.html

Most artificial neural networks have been designed based on computer software which is often limited by Von Neumann's bottleneck. However due to the development of memristor materials there may be a move to create more hardware-based neural architectures. This patent from Elpida Memory, a maker of DRAM chips, is a variation of such neuromorphic hardware based on phase change material. Claim 1 reads:

1. An artificial neural network comprising:

a learning circuit that stores an input history of an input signal; and

an output circuit that is connected to the learning circuit, wherein the learning circuit changes a potential-change characteristic of an internal node included in the output circuit according to the input history, and the output circuit starts an output operation of data when a potential at the internal node exceeds a threshold value.

US Patent 7542951 - Diverse recommendations

http://www.freepatentsonline.com/7542951.html

Amazon uses an automated recommendation algorithm to suggest new items for purchase to returning customers based on their purchasing history. This patent teaches a method that provides some less obvious recommendations and generates a greater diversity of suggestions. Claim 1 reads:

1. A method for generating diverse recommendations, the method implemented at least in part by a computing device and comprising:

providing at least one source of information from which to generate recommendations;

generating a set of original recommendations based on said at least one source of information;

generating a set of diverse recommendations from the set of original recommendations based on repulsion force analysis applied to the set of original recommendations, wherein the repulsion force analysis ensures that at least one recommendation is separated from a reference by at least a sufficient amount to provide a desired degree of variation between said at least one recommendation and the reference; and

providing the set of diverse recommendations to a user, wherein the reference is another recommendation, such that the repulsion force analysis ensures that said at least one recommendation is separated from the other recommendation by the sufficient amount.

US Patent 7530410 - Robotic toe

http://www.freepatentsonline.com/7530410.html

Proper posture and balance are a significant problem for humaniform robots. This patent from Honda teaches one solution based on an improved design for robotic toes.

1. A legged mobile robot having a body and legs whose upper ends are connected to the body and whose lower ends are each connected to a foot to be movable when the legs are driven, wherein the foot comprises:

a foot main body connected to each of the legs;

a toe provided at a fore end of the foot main body to be bendable with respect to the foot main body; and

a bending angle holder capable of holding a bending angle of the toe in a bendable range from zero degree at which a contact area of the foot becomes maximum to a predetermined degree, the bending angle holder holding the toe at the bending angle comprising one from among the zero degree, the predetermined degree and an arbitrary angle between the zero degree and the predetermined angle.

US Patent 7526455 - Decision support system

http://www.freepatentsonline.com/7526455.html

This is another patent from Palomar Technology including some very basic claims to a computer based system for making a recommendation based on a variety of data source inputs. Claim 1 reads:

1. A decision support system, comprising:

a database containing data from a plurality of disparate data sources; and

an electronic device configured to classify the database data and assign weighted values to the data based at least in part on the classification, the electronic device further configured to analyze the database data so as to provide at least one recommended action and outcome information for the at least one recommended action.

However, some prior art internet services would seem to anticipate at least some of the claims such as US 2005/0193002 teaching an internet recommendation engine from Yahoo.

US Patent 7523080 - Self-organizing model for artificial life

US Patent 7523080

This patent is interesting in that it attempts to claim a robotic system architecture and A.I. replicating the instinctive behavior of living organisms for survival and decision making including both conscious and sub-conscious for the robotic system. Claim 1 reads:

1. An architecture for a self-organizing robotic system comprising:

an input system which comprises:

sensory organ means for receiving information from the environment and translating the information into an information pattern; and

input receptive organ means for receiving each information pattern and converting each information pattern into a geometric connection;

an output system which comprises:

output receptive organ means for receiving geometric connections; and

effect organ means for receiving the geometric connections from the output receptive organ means and for causing an effect in response;

a mind system for receiving geometric connections from the input system, processing the geometric connections and transmitting the geometric connections to the output system, which mind system comprises:

an advanced mind subsystem comprising an advanced neuro-matrix means for providing consciousness; and a primitive mind subsystem comprising:

a primitive neuro-matrix means for providing subconsciousness;

instinct means for storing causal relationships to provide a geometric connection output according to a geometric connection input;

instinct for survival means for prioritizing each input and output in terms of comfortable and uncomfortable; and

instincts arbiter means for controlling future actions through promotion and inhibition of the output geometric connection according to status of the instinct for survival means;

wherein the output geometric connection observes conditions expressed by the instinct for survival means and the instincts arbiter means.

US Patent 7519564 - Predictive model of surprising events

http://www.freepatentsonline.com/7519564.html

This patent from Microsoft teaches applying machine learning toward the prediction of future events. Claim 1 reads:

1. A system that predicts and outputs events identified as being surprising to a person, comprising a memory having stored therein computer executable components and a processor that executes the following computer executable components:

an interface component that receives contextual and historical data;

a predictive model component that utilizes the contextual and historical data to predict an event and outputs the prediction if the prediction corresponds to one or more definitions of surprise, wherein the prediction corresponds to one or more definitions of surprise based on a probability of occurrence of the event, and wherein the predictive model component comprises:

a robust predictive model that generates a prediction of the event based on interdependencies between variables associated with the contextual and historical data, wherein the interdependencies are not contemplated by the person; and

a user expectancy model that utilizes the contextual and historical data to generate a prediction of the event based on, at least in part, machine learning and a case library that includes a plurality of surprising events and observations associated with the plurality of surprising events;

a difference analyzer component that calculates a measure of difference between the prediction made by the robust predictive model and the prediction made by the user expectancy model to determine whether an event is surprising; and

an alerting component that alerts the person of the surprising event upon the determination that the event is surprising.