Vitaly Feldman, 46Mountain View, CA

2002009, Jul 18, 2002

Mar 16, 2001

09/727942

Vitaly Feldman - Newton MA, US
Sergio Manevich - Winchester MA, US

B26D003/00
B26D005/00

083/049000, 083/056000, 083/076900, 083/075500

A method and system of cutting parts from sheet material comprising a numerically-controlled cutting system having a cutting tool for cutting along a path, the method comprising the steps of: (a) placing a plurality of templates which define the shapes and sizes of the parts to be cut upon the sheet material while minimizing the spaces between the templates to form a closely-packed marker; (b) entering the marker into a pre-processor; (c) detecting common lines and tangencies between templates in the marker; (d) determining a path and speed for said cutting tool; and (e) directing the cutting tool in accordance with the path and speed such that the parts are cut from the sheet material. The pre-processor identifies critical segments of the cutting path where cutting difficulties may arise and modifies the data that guide the cutting tool for more accurate cutting through the critical segments. In particular, the pre-processor identifies segments of the cutting path proximately close to one another called “common line segments” and generates a modified cutting path using a single pass to cut common line segments. The method and system of the present invention provide an optimal cutting path and control of a cutting tool resulting in higher quality cut pieces and the highest possible throughput. Therefore, the method disclosed allows for and makes desirable the close nesting of templates without buffers.

2013031, Nov 28, 2013

May 25, 2012

13/480974

Laura Chiticariu - San Jose CA, US
Vitaly Feldman - Palo Alto CA, US
Frederick R. Reiss - Sunnyvale CA, US
Sudeepa Roy - , US
Huaiyu Zhu - Union City CA, US

INTERNATIONAL BUSINESS MACHINES CORPORATION - Armonk NY

G06F 17/30

707723, 707E17014

A method for refining a dictionary for information extraction, the operations including: inputting a set of extracted results from execution of an extractor comprising the dictionary on a collection of text, wherein the extracted results are labeled as correct results or incorrect results; processing the extracted results using an algorithm configured to set a score of the extractor above a score threshold, wherein the score threshold balances a precision and a recall of the extractor; and outputting a set of candidate dictionary entries corresponding to a full set of dictionary entries, wherein the candidate dictionary entries are candidates to be removed from the dictionary based on the extracted results.

2013031, Nov 28, 2013

Aug 30, 2012

13/598946

Laura Chiticariu - San Jose CA, US
Vitaly Feldman - Palo Alto CA, US
Frederick R. Reiss - Sunnyvale CA, US
Sudeepa Roy - , US
Huaiyu Zhu - Union City CA, US

International Business Machines Corporation - Armonk NY

G06F 17/30

707723, 707E17014

A method for refining a dictionary for information extraction, the operations including: inputting a set of extracted results from execution of an extractor comprising the dictionary on a collection of text, wherein the extracted results are labeled as correct results or incorrect results; processing the extracted results using an algorithm configured to set a score of the extractor above a score threshold, wherein the score threshold balances a precision and a recall of the extractor; and outputting a set of candidate dictionary entries corresponding to a full set of dictionary entries, wherein the candidate dictionary entries are candidates to be removed from the dictionary based on the extracted results.

2021015, May 27, 2021

Nov 20, 2020

16/953977

- Mountain View CA, US
Vitaly Feldman - Mountain View CA, US
Tomer Koren - Herzliya, IL

G06N 20/00
G06F 17/18
G06F 21/60

Methods, systems, and apparatus, including computer programs encoded on computer storage media for training a machine learning model. The method includes obtaining a training data set comprising a plurality of training examples; determining i) a stochastic gradient descent step size schedule, ii) a stochastic gradient descent noise schedule, and iii) a stochastic gradient descent batch size schedule, wherein the stochastic gradient descent batch size schedule comprises a sequence of varying batch sizes; and training a machine learning model on the training data set, comprising performing stochastic gradient descent according to the i) stochastic gradient descent step size schedule, ii) stochastic gradient descent noise schedule, and iii) stochastic gradient descent batch size schedule to adjust a machine learning model loss function.