What is claimed is:
1. A method of mapping a set of input patterns to an m-dimensional space, comprising the steps of:
(a) selecting k patterns from said set of input patterns to form a subset of patternspi, i 1, . . . , k;
(b) determining at least some pairwise relationships between at least some of the patterns in said subset of patterns pi;
(c) mapping the patterns pi into a set of images in an m-dimensional space (piyi, i1, 2, . . . k, yi Rm) so that at least some of the pairwise distances between at least some of the images yi are representative of the relationships of the respective patterns pi;
(d) determining a set of n attributes for each pattern in said subset of patterns pi, xi, i1, 2, . . . k, xi Rn;
(e) forming a training set T(xi, yi), i1, 2, . . . k;
(f) using a supervised machine learning technique to determine a mapping function based on the training set T; and
(g) using said mapping function determined in step (f) to map additional patterns.
2. The method of claim 1, wherein said mapping function is encoded in at least one neural network.
3. The method of claim 1, wherein step (f) further comprises the steps of:
(i) determining c n-dimensional reference points, ci, i1, 2, . . . c, ci Rn;
(ii) partitioning T into c disjoint clusters Cj based on a distance function d, Cj(xi, yi): d(xi, cj).d(xi, ck) for all kj; j1, 2, . . . c; i1,2, . . . k; and
(iii) using a supervised machine learning technique to determine c independent mapping functions fiL, i1, 2, . . . c, based on the respective subsets Ci of the training set T.
4. The method of claim 3, wherein each said mapping function is encoded in at least one neural network.
5. The method of claim 3, wherein step (g) further comprises the steps of:
(i) for an additional input pattern p, determining a set of n attributes x, x Rn;
(ii) determining the distance of x from each said reference point ci;
(iii) identifying the reference point cj closest to x; and
(iv) mapping xy Rm, using the mapping function fjL associated with the reference point cj identified in step (iii).
6. The method of claim 5, wherein each said mapping function is encoded in at least one neural network.
7. The method of claim 3, wherein step (i) is performed using a clustering algorithm.
8. The method of claim 1, wherein step (f) further comprises the steps of:
(i) determining c m-dimensional reference points, ci, i1, 2, . . . c, ci Rm;
(ii) partitioning T into c disjoint clusters Cj based on a distance function d, Cj(xi, yi): d(yi, cj)d(yi, ck) for all kj; j1, 2, . . . c; i1,2, . . . k;
(iii) using a supervised machine learning technique to determine c independent local mapping functions fiL, i1, 2, . . . c based on the respective subsets Ci of the training set T; and
(iv) using a supervised machine learning technique to determine a global mapping functionf, based on the entire training set T.
9. The method of claim 8, wherein each said mapping function is encoded in at least one neural network.
10. The method of claim 8, wherein step (g) further comprises the steps of:
(i) for an additional input pattern p, determining a set of n attributes x, x Rn;
(ii) mapping xy, y Rm, using the global mapping function
(iii) determining the distance of y to each reference point in ci;
(iv) identifying the reference point cj closest to y; and
(v) mapping xy, y Rm, using the local mapping function fjL associated with the reference point cj identified in step (iv).
11. The method of claim 10, wherein each said mapping function is encoded in at least one neural network.
12. The method of claim 8, wherein step (i) is performed using a clustering algorithm.
13. The method of claim 1, wherein step (c) further comprises the steps of:
(i) selecting a subset of patterns from pi;
(ii) revising the positions of the images of said selected subset of patterns in the m-dimensional space based on the relationships between said selected subset of patterns determined in step (b); and
(iii) repeating steps (i) and (ii) for additional subsets of patterns from pi.
14. The method of claim 1, wherein step (d) comprises the step of:
(i) determining a set of n attributes for each pattern in said subset of patterns pi, xi, i1, 2, . . . k, xi Rn, wherein said attributes represent the relationships of each pattern in said subset of patterns with respect to n other reference patterns.
15. The method of claim 1, wherein step (b) comprises the step of receiving pairwise relationship data from a subject.
16. The method of claim 15, wherein step (b) comprises the steps of:
(1) randomly selecting two patterns from the plurality of patterns;
(2) presenting the two patterns to at least one subject; and
(3) receiving pairwise relationship data from said subject about the patterns.
17. The method of claim 16, further comprising the step of:
(4) repeating steps (b)(1) through (b)(3) for additional pairs of patterns.
18. The method of claim 17, wherein step (b) further comprises the step of receiving data about the subjects providing the pairwise relationship data.
19. The method of claim 1, wherein step (b) comprises the steps of:
(1) receiving pairwise relationship data via a communications path coupled to a computer; and
(2) storing the received pairwise relationship data in a memory.
20. The method of claim 1, wherein step (b) comprises the steps of:
(1) selecting a pair of patterns for similarity comparison;
(2) transmitting information about the selected pair of patterns to a remote computer system; and
(3) receiving pairwise relationship data about the selected pair of patterns from the remote computer system.
21. The method of claim 20, further comprising the step of:
(4) repeating steps (b)(1) through (b)(3) for additional pairs of compounds.
22. The method of claim 21, wherein step (b)(1) comprises the step of selecting the pair of patterns at random.
23. A computer program product comprising a computer useable medium having computer program logic recorded thereon for enabling a processor to obtain pairwise relationship data about the selected pair of patterns, said computer program logic comprising:
a selecting procedure that enables the processor to select a plurality of patterns from a database for similarity;
a transmitting procedure that enables a processor to transmit selected patterns to a remote computer for similarity comparison; and
a receiving procedure that enables the processor to receiving similarity data about patterns transmitted to the remote computer.
24. The computer program product of claim 23, further comprising:
a storing procedure that enables the processor to store received similarity data for subsequent retrieval.
25. The computer program product of claim 24, wherein said selecting procedure randomly selecting patterns from the database.
26. A system for mapping a set of input patterns to an m-dimensional space, comprising:
means for selecting k patterns from said set of input patterns to form a subset of patterns pi, i1, . . . , k;
means for determining at least some pairwise relationships between at least some of the patterns in said subset of patterns pi;
means for mapping the patterns pi into a set of images in an m-dimensional space (piyi, i1, 2. . . k, yi Rm) so that at least some of the pairwise distances between at least some of the images yi are representative of the relationships of the respective patterns pi;
means for determining a set of n attributes for each pattern in said subset of patterns pi, xi, i1, 2, . . . k, xi Rn;
means for forming a training set T(xi, yi), i1, 2, . . . k;
means for using a supervised machine learning technique to determine a mapping function based on the training set T; and
means for using said mapping function determined in step (f) to map additional patterns.
27. A computer program product comprising a computer useable medium having computer program logic recorded thereon for enabling a processor to obtain map a set of input patters to an m-dimensional space, said computer program logic comprising:
a procedure that selects k patterns from said set of input patterns to form a subset of patterns pi, i1, . . . , k;
a procedure that determines at least some pairwise relationships between at least some of the patterns in said subset of patterns pi;
a procedure that maps the patterns pi into a set of images in an m-dimensional space (piyi, i1, 2, . . . k, yi Rm) so that at least some of the pairwise distances between at least some of the images yi are representative of the relationships of the respective patterns pi, ;
a procedure that determines a set of n attributes for each pattern in said subset of patterns pi, xi, i1, 2, . . . k, xi Rn;
a procedure that forms a training set T(xi, yi), i1, 2, . . . k ;
a procedure that uses a supervised machine learning technique to determine a mapping function based on the training set T; and
a procedure that uses said mapping function determined in step (f) to map additional patterns.
The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.
1. A tool coupler system for a machine, comprising:
a tool coupler having a hydraulic actuator configured to selectively lock a tool to the machine;
a first hydraulic pump configured to selectively generate a first flow of fluid pressurized to a first pressure and supply the fluid pressurized to the first pressure to the hydraulic actuator of the tool coupler and at least one other hydraulic actuator, wherein the first hydraulic pump is configured to supply the first flow of fluid pressurized to the first pressure to the hydraulic actuator of the tool coupler to lock the tool to the machine;
a second hydraulic pump configured to selectively generate a second flow of fluid pressurized to a second pressure and supply the fluid pressurized to the second pressure to the hydraulic actuator of the tool coupler, wherein the second hydraulic pump is configured to supply the second flow of fluid pressurized to the second pressure to the hydraulic actuator of the tool coupler to maintain the tool coupler in the locked position, wherein the second flow of fluid pressurized to the second pressure is less than an amount required to move the hydraulic actuator of the tool coupler to either lock or unlock the tool from the machine; and
a shuttle valve configured to selectively move to a first position when the first pressure is greater than the second pressure to allow the first flow of fluid pressurized to the first pressure to flow through the shuttle valve in a first direction to supply pressurized fluid from the first hydraulic pump to the hydraulic actuator of the tool coupler, or to a second position when the second pressure is greater than the first pressure to allow the second flow of fluid pressurized to the second pressure to flow through the shuttle valve in a second direction opposite the first direction to supply pressurized fluid from the second hydraulic pump to the hydraulic actuator of the tool coupler.
2. The tool coupler system of claim 1, wherein both the first and second pumps are variable displacement pumps.
3. The tool coupler system of claim 2, wherein both the first and second pumps are driven by an engine of the machine.
4. The tool coupler system of claim 1, wherein the first hydraulic pump is an implement pump of the machine.
5. The tool coupler system of claim 4, wherein the first pressure of the first flow of pressurized fluid is driven by a demand for pressurized fluid from the at least one other hydraulic actuator.
6. The tool coupler system of claim 5, wherein the shuttle valve directs the second flow of fluid pressurized to the second pressure to the hydraulic actuator of the tool coupler only when the at least one other hydraulic actuator is idle.
7. The tool coupler system of claim 6, wherein the first hydraulic pump is destroked to a neutral position when the at least one other hydraulic actuator is idle.
8. The tool coupler system of claim 1, wherein the second hydraulic pump is a pilot pump configured to supply pilot fluid to move at least one valve of the machine.
9. The tool coupler system of claim 8, wherein the first pressure is about 10 times the second pressure when both the first and second pumps are pressurizing fluid.
10. The tool coupler system of claim 1, wherein:
the tool coupler system further includes a second valve configured to control a flow direction of pressurized fluid through the hydraulic actuator of the tool coupler.
11. The tool coupler system of claim 1, wherein:
the tool coupler further includes:
a coupler frame;
a hook configured to receive a first pin of the tool; and
a wedge;
the first hydraulic pump is configured to direct the first flow of pressurized fluid to the hydraulic actuator of the tool coupler to move the wedge away from the hook and bias a second pin of the tool against the coupler frame; and
the second hydraulic pump is configured to direct the second flow of pressurized fluid to the hydraulic actuator of the tool coupler to maintain the wedge away from the hook.
12. The tool coupler of claim 11, further including a check valve configured to maintain fluid having a pressure about the same as the first flow of pressurized fluid within a head-end chamber of the hydraulic actuator of the tool coupler even when the second flow of pressurized fluid is being directed to the hydraulic actuator of the tool coupler.