Recommind, Inc.

Open Text Holdings, Inc.

Jan Puzicha

Thomas Hofmann

Steve Vranas

Jan Stadermann

Sangeetha Yanamandra

Dietrich Wettschereck

Ketan Deshpande

Joe Federline

Chaitanya Muppala

Christian Feuersänger

Robert Tennant

Electronic discovery using predictive filtering is disclosed herein. An example method includes providing a filtering interface that includes a field value input, a predicted values selector, and a predictor type selector; receiving at least a pivot selected from the field value input and a predicted value from the predicted values selector; searching a plurality of documents based on the pivot and the predicted value selected for any of predictive phrases or predictive concepts; calculating a predictive value for each of the predictive phrases or predictive concepts; and generating a graphical user interface that includes the predictive phrases or predictive concepts in conjunction with their respective predictive value.

Systems and methods of predictive filtering using document field values

Systems and methods for analyzing documents are provided herein. A plurality of documents and user input are received via a computing device. The user input includes hard coding of a subset of the plurality of documents, based on an identified subject or category. Instructions stored in memory are executed by a processor to generate an initial control set, analyze the initial control set to determine at least one seed set parameter, automatically code a first portion of the plurality of documents based on the initial control set and the seed set parameter associated with the identified subject or category, analyze the first portion of the plurality of documents by applying an adaptive identification cycle, and retrieve a second portion of the plurality of documents based on a result of the application of the adaptive identification cycle test on the first portion of the plurality of documents.

Systems and methods for predictive coding

Provided herein are systems and methods for user-defined automated document feature modeling, extraction and optimization. In the present disclosure, an end user of an automated document review system can customize and create new data models applicable to a set of focus documents. In addition, an end user of the automated document review system can customize and create new extraction rules applicable to text extraction from the set of focus documents. The user-defined edits to the data model and extraction rules can be further tested in a staging environment, and tested against a ground truth set of documents, before being widely applied to other relevant documents.

User-defined automated document feature modeling, extraction and optimization

Electronic discovery using predictive filtering

Provided herein are systems and methods for user-defined automated document feature modeling, extraction and optimization. In the present disclosure, an end user of an automated document review system can customize and create new extractor taggers within data models applicable to a set of focus documents. The user-defined edits to the extractor taggers can be further tested in a staging environment, and tested against a ground truth set of documents, before being widely applied to other relevant documents.

User-defined automated document feature extraction and optimization

Systems and methods that quantify document relevance for a document relative to a training corpus and select a best match or best matches are provided herein. Methods may include generating an example-based explanation for relevancy of a document to a training corpus by executing a support vector machine classifier, the support vector machine classifier performing a centroid classification of a relevant document in a term frequency-inverse document frequency features space relative to training examples in a training corpus, and generating an example-based explanation by selecting a best match for the relevant document from the training examples based upon the centroid classification.

Document relevancy analysis within machine learning systems

Methods and systems of integrated batching and random sampling of documents for enhanced functionality and quality control, such as validation, within a document review process are provided herein. According to various embodiments, a batching request may be received and may include a population size that corresponds to a total amount of documents available for sampling. The batching request may also include an acceptable margin of error. A random sample size may be calculated based on the batching request, and then a subset of documents corresponding to the random sample size may be selected from the total amount of documents available for sampling. The subset of documents may be grouped into one or more batches, and the one or more batches may be assigned to one or more review nodes.

Integration and combination of random sampling and document batching

A computer readable storage medium includes executable instructions to display a document with legal categorizations. The legal categorizations are accepted and invoke a subsequent document for viewing with a single input action. The subsequent document is displayed. Revised legal categorizations are received. The revised legal categorizations are accepted and invoke an additional document for viewing with a single input action.

Apparatus and method for single action approval of legally categorized documents

A system and method for information navigation and filtration is provided. One or more query terms are received from a user. A preliminary relevance of one or more objects associated with an enterprise system is determined based on the query terms. The preliminary relevance may be propagated between objects. At least one rating is assigned to the one or more objects based on the preliminary relevance. An overall relevance of the one or more objects is established based on the at least one rating. The one or more objects are ranked according to the overall relevance. Data is provided as search results comprised of the one or more objects according to the ranking to the user. The search results may then be filtered based on at least one selected, dynamically generated filter. The filtered search results may be dynamically generated and provided to the user.

System and method for providing information navigation and filtration

A system, method, and computer program for establishing relevance of objects in an enterprise system is provided. One or more objects are assigned to content associated with an enterprise system. One or more query terms are received from a user. A preliminary relevance of the one or more objects is determined based on the query terms. Ratings are assigned to the one or more objects based on the preliminary relevance. An overall relevance of the one or more objects is established based on the ratings.

System and method for establishing relevance of objects in an enterprise system

Systems and methods that quantify document relevance for a document relative to a training corpus and select a best match or best matches are provided herein. Methods may include generating an example-based explanation for relevancy of a document to a training corpus by executing a support vector machine classifier, the support vector machine classifier performing a centroid classification of a relevant document in a term frequency-inverse document frequency features space relative to training examples in a training corpus, and generating an example-based explanation by selecting a best match for the relevant document from the training examples based upon the centroid classification. Determining the training example having the closest cosine distance to the relevant document includes ranking the training examples by stretching the internal best match scores for the training examples linearly to cover a complete unit interval.

Growing community of inventors

Bonn, Germany

Jan Puzicha