US20040163035A1 - Method for automatic and semi-automatic classification and clustering of non-deterministic texts - Google Patents
Method for automatic and semi-automatic classification and clustering of non-deterministic texts Download PDFInfo
- Publication number
- US20040163035A1 US20040163035A1 US10/771,315 US77131504A US2004163035A1 US 20040163035 A1 US20040163035 A1 US 20040163035A1 US 77131504 A US77131504 A US 77131504A US 2004163035 A1 US2004163035 A1 US 2004163035A1
- Authority
- US
- United States
- Prior art keywords
- recited
- documents
- word sequences
- data mining
- word
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
- G10L15/183—Speech classification or search using natural language modelling using context dependencies, e.g. language models
- G10L15/19—Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
- G10L15/197—Probabilistic grammars, e.g. word n-grams
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/38—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/38—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/383—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9538—Presentation of query results
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/98—Detection or correction of errors, e.g. by rescanning the pattern or by human intervention; Evaluation of the quality of the acquired patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/24—Character recognition characterised by the processing or recognition method
- G06V30/248—Character recognition characterised by the processing or recognition method involving plural approaches, e.g. verification by template match; Resolving confusion among similar patterns, e.g. "O" versus "Q"
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
Definitions
- the present invention is directed to processing of information in non-deterministic texts to increase the usefulness of the texts and, more particularly, to using non-textual information to indicate the importance or recognition accuracy of individual words or sequences of words.
- spoken document retrieval is composed of two stages: transcription of speech and information retrieval (IR). Transcription of the speech is often referred to as speech-to-text (STT) or automatic speech recognition (ASR), and is often performed using a large vocabulary continuous speech recognizer (LVCSR).
- Information retrieval is a general term referring to all forms of data mining.
- One common form of data mining for example, is query-based retrieval, where, based on a user's query, documents are retrieved and presented to the user, ordered by an estimated measure of their relevance to the query. Traditionally, this stage is performed on the text output of the first stage.
- the above aspects can be attained by a method for processing documents derived from at least one of spontaneous and conversational expression and containing non-deterministic text with average word recognition precision below 50 percent, the processing utilizing non-textual differences between n-word sequences in the documents to resolve more than two decision options, where n is a positive integer.
- Such text may be obtained by automatic character recognition or automatic speech recognition of audio signals received via a telephone system.
- the non-textual differences between the n-word sequences relate to recognition confidence of the n-word sequences
- the processing requires fast access to the information stored in a large corpus of documents, e.g. for the purpose of data mining, the data is preferrably pre-processed to index the n-word sequences in a method that utilizes the non-textual differences between them.
- Such a procedure can speed up many forms of data access, and in particular many forms of data mining, including query based retrieval, as would be apparent to a person skilled in the art.
- the data mining may include extracting parameters from the documents utilizing the non-textual differences between the n-word sequences and establishing relations between the parameters extracted from the documents.
- the parameters extracted from the documents may be fully known, such as parameters available in document metadata or may be hidden variables that cannot be fully determined from information existing in the document. Examples of extracted parameters include an assessment of relevance to a query based on the non-textual differences between the n-word sequences and an assessment of the document's relevance to a category.
- algorithm(s) can be used to convert text containing non-textual differences between the n-word sequences into different standard text documents.
- Many different algorithms may be used to transform non-deterministic text into standard text documents usable in text mining.
- the algorithm to extract standard text documents from text with non-textual differences may apply a thresholding algorithm with varying thresholds.
- one or more data mining techniques, each of which does not utilize non-textual differences can be applied to these standard text documents and the outputs of the different data mining techniques can be merged to obtain information that is equivalent to that obtained by data mining that utilizes the non-textual differences.
- the documents may be categorized, clustered, classified, filtered or labeled, e.g., by using an algorithm to detect salient terms in the documents based on non-linguistic differences between the n-word sequences.
- information related to at least one of the documents may be displayed, including at least some non-textual differences between n-word sequences.
- Portions of the document(s) may be selectively displayed based on confidence of the accuracy of the displayed words. For example, salient terms in the document(s) may be displayed based on processing of confidence levels of recognition of the salient terms that resolves more than two decision options.
- parameters extracted from the documents and indications of the relations between these parameters may be displayed graphically.
- a user may indicate errors in recognition.
- at least one word in the document is preferably replaced with a corrected word supplied by the user and the confidence level(s) of the corrected word(s) are reset to indicate high recognition accuracy.
- FIG. 1 is a flowchart of a conventional spoken document retrieval system.
- FIG. 2 is a flowchart of one method of spoken document retrieval according to the invention.
- FIG. 3 is a flowchart of another method of spoken document retrieval according to the invention.
- FIGS. 4 and 5 are block diagrams of spoken document retrieval systems according to the invention.
- FIG. 6 is a block diagram of one confidence sensitive inverted index and one regular index containing confidence information.
- FIG. 7 is a flowchart of text processing according to the invention.
- FIGS. 8 and 9 are examples of displays generated by telephone call processing applications according to the invention.
- labeling a form of data processing where documents are analyzed, and the analysis results (referred to as “labels”) are made available for later processing stages.
- a topical analysis of documents is a labeling of the documents by subject.
- retrieving a form of data mining where a subset of a document corpus is returned in response to a query.
- the documents are each given a rank pertaining to their relevance to the query, and are sorted by decreasing relevance. categorizing a form of data mining where several “categories” are defined, and the documents of a corpus are labeled according the category to which they fit best.
- a common variation is multilabel categorizing, where each document may fit zero or more categories. Preferably, information is given regarding the quality of the fit.
- clustering a form of data mining similar to categorization, with the difference that the “categories” are not predefined, and the data mining must reveal them automatically.
- classifying a process performed on a stream of incoming documents, where each is labeled and then forwarded for relevant additional processing (manual or automatic) based on the labels that have been discovered.
- filtering a process performed on a stream of incoming documents, where each is labeled and then forwarded or discarded based on the labels that have been discovered.
- salient terms terms whose appearance in a document provides information relevant to its correct labeling, and consequently to all forms of data mining subsequent to labeling.
- High quality audio 20 is input into an ASR system 22 using LVCSR.
- ASR system 22 converts spoken words into textually represented words, but often has other outputs as well. These outputs may include timing information, an indication of the confidence of recognition of particular words and phrases, alternative likely transcriptions, and more.
- the LVCSR output cannot be piped directly into a traditional text mining system. It has to be converted into searchable text. For this reason, canonization 24 is performed to produce canonized text, also referred to below as standard text documents, used by conventional text mining software. Most commonly, canonization simply involves taking the textual words out of the LVCSR output and concatenating them. More sophisticated canonization schemes involve usage of both textual and non-textual information to convert the LVCSR output into a format more easily handled by text mining system 26 . Usage of textual information may include capitalization and punctuation based on grammatical rules. Usage of non-textual information may include capitalization and punctuation using timing information and word(s) omitted based on low confidence levels.
- Text mining system 26 receives input from many different audio segments and stores the information in some format that will be convenient for later processing, a process known as “indexing”. When asked to produce output, typically, though not exclusively, by user query 28 , text mining system 26 searches its index and produces output. For example the output may be the identities of the audio segments that were requested for retrieval, ranked and scored by some relevance metric. The output may also include other information, such as the phrases in the retrieved segments that have proved to be salient terms, because of which the document was given the score that it was given.
- document display system 30 can use all of it, and add to it the original audio segment(s) 20 , to give the user feedback to user query 28 that is as informative and audio-visually appealing as possible.
- FIG. 2 A simplified embodiment of a method according to the present invention for spoken document retrieval of low grade audio 32 is illustrated in FIG. 2.
- a similar system could be used for text generated by a handwritten text recognition system.
- the traditional text mining system has been replaced with data mining system 34 that is designed especially for speech in low grade audio 32 .
- Data mining system 34 doesn't require that the output of ASR 36 be canonized into text before it is handled, and can therefore utilize all information available in the output of ASR 36 . No information is lost in a canonization process, and words that are indexed can receive different and appropriate handling based on non-textual information, such as their confidence scores.
- ASR 36 is therefore entirely omitted, and the output of ASR 36 is available with more of its information in subsequent processing stages, including indexing, retrieval and display. Since non-textual information is available, document display system 38 in FIG. 2 displays different information than document display system 30 in FIG. 1.
- Data mining system 34 uses data mining targeted specifically at low-grade audio speech data, extracting useful information by ranking and retrieval utilizing non-textual data which makes data mining of telephony speech quite feasible.
- data mining 34 includes indexing which differentiates words using non-textual differences between the words.
- indexing which differentiates words using non-textual differences between the words.
- data mining system 34 uses the non-textual differences between the words for retrieval.
- FIG. 3 An alternative way to determine non-textual differences between words without using a specially created data mining system is illustrated in FIG. 3.
- the same segments of audio 32 are canonized several times using different canonization methods 24 a , . . . 24 n .
- These methods can differ, for example, by a choice of minimal allowed confidence, any word of confidence lower than that being removed from the transcription in one way or another.
- the obtained text is indexed in N different indexes by text mining systems 26 a , . . . 26 n which may use the same indexing methods or different methods.
- low grade audio 32 may additionally be processed by ASR 36 to produce output for data mining system(s) 34 that utilize(s) non-textual differences between words.
- text mining systems 26 a , . . . 26 n search indexes of standard text documents, result(s) are retrieved from each index, and finally all N results are algorithmically merged 42 , providing a single output, which is then forwarded to document display system 38 .
- the merged output unlike the single outputs of the text mining systems, can differentiate results by confidence levels, and display system 38 can use this information.
- FIGS. 4 and 5 Either of the spoken document retrieval methods illustrated in FIGS. 2 and 3 can be implemented using systems like that illustrated in FIGS. 4 and 5. Such systems can be configured in many different ways, depending on the tasks needed to be performed and the volume of data processed and FIGS. 4 and 5 are just two examples. The invention is not limited to the configurations illustrated in FIGS. 4 and 5 and other configurations are possible and will be apparent to a person of ordinary skill in the art. For example, the functions performed by separate servers in FIGS. 4 and 5 may be performed by separate modules in a single computing system.
- Voice data are supplied from voice acquisition module(s) 50 by network 52 and stored in data storage 54 .
- Network 52 may be any known type of network, such as a local area network (LAN), wide area network (WAN), the Internet, etc.
- the voice data may be in the form of WAV files or any other audio file format.
- the system is accessed by one or more user terminals 56 , such as personal computers or other devices that include a user interface which may include a display.
- user terminals 56 such as personal computers or other devices that include a user interface which may include a display.
- users log into the system at various times to submit queries to voice oriented information retrieval (VOIR) indexing server 58 .
- Voice data from voice acquisition module(s) 50 are supplied to speech categorization server 60 which, if necessary, converts the data before supplying the voice data to LVCSR(s) 22 and performs load balancing when more than one LVCSR 22 is used.
- LVCSR(s) 22 output words and additional data, such as speaker-change, timing information, confidence scores, etc.
- call metadata such as the time that a call was made and the number dialed, is obtained from voice acquisition module(s) 50 together with the voice data. All these types of data are combined, e.g., by speech categorization server 60 and forwarded, in online mode to speech analysis server 62 and in offline mode to VOIR indexing server 58 . Regardless of whether the method illustrated in FIG. 2 or FIG. 3 is implemented, results of a query in offline mode can be displayed on user terminals 56 with at least some of the non-textual differences between n-word sequences indicated. Examples of how the non-textual differences are conveyed to the user will be described below with reference to FIGS. 7 and 8.
- the online configuration illustrated in FIG. 5, may be used when the volume of voice data is too large to allow effective offline processing, or it is desired to use push-technology alerts to people who may want the data. For example, a police inspector may want to be paged when the system detects a phone conversation relevant to her case.
- the output of LVCSR(s) 22 is supplied via network 52 to speech analysis server 62 which labels the voice data.
- the voice data may be labeled according to importance, subject matter, person or group that needs to respond, etc.
- the labeling of the transcribed voice data is combined with the output of LVCSRs 22 and call metadata, and forwarded to categorization queue and workflow manager 64 .
- categorization queue and workflow manager 64 The users at user terminals 56 are provided this information by categorization queue and workflow manager 64 .
- categorization queue and workflow manager 64 uses the labeling provided by speech analysis server 62 to categorization queue and workflow manager 64 supplies text, voice data and call metadata appropriate for that user, depending on importance, topic, identity of the user, etc.
- Training of speech analysis server 62 may be accomplished by offline processing using VOIR indexing server 58 in an implementation that includes both servers 58 and 62 .
- One or more users label calls by importance, subject matter, relevant person or group, etc.
- the labels assigned by users can be provided to speech analysis server 62 as training data to recognize similar calls during online processing of calls in a call center, for example.
- training may continue during online processing as users correct the labeling provided by speech analysis server 62 .
- VOIR indexing server 58 is trained in a similar manner.
- LVCSR 22 supplies metadata, including confidence scores, associated with recognized words to VOIR indexing server 58 which generates an index that indicates at least some of the non-textual differences between n-word sequences. If the method described above with reference to FIG. 3 is implemented, VOIR indexing server 58 maintains an index for each canonization system 24 . In either case, the index(es) and the voice data (preferably compressed to minimize space requirements) or other data from which indexed text is obtained (such as handwritten documents) are preferably stored in data storage 54 .
- data in data storage 54 is indexed by use of at least one confidence sensitive inverted index.
- a confidence sensitive inverted index maps from terms to a sorted linked list identifying all documents where each term occurs and from each appearance of a document in this list to a sorted linked list identifying all positions in which the term appears and the confidence level of its recognition.
- indexed data may include aggregated information relating to confidence.
- each appearance 69 of a term in the document can carry additional data, such as its position in the document, its timing information, recognition score of that appearance of the term, etc.
- an expected number of real occurrences of the term e.g., term i which points to mapping 67 i
- the indexed document e.g., 68 a
- Another example of aggregated information relating to confidence information that can be saved is the strength of association between every document and each category.
- This information can be saved either in a regular (forward) index, like index 66 , another inverted index (not shown), or both.
- Information not relating to confidence such as call metadata, can also be indexed, either in another inverted index, in a forward index like index 66 , or both. In both cases, if an inverted index is used, confidence sensitive inverted index 65 or a separate index can be used.
- additional mapping technologies in addition to or instead of a mapping into a sorted linked list, can also be used.
- Data storage 54 can also store information other than the indexes, such as the data that is being indexed. This data may include, among others, call audio, voice data and call metadata, and may include additional indexes used to refer to the same data.
- FIG. 8 A more detailed flow of processing through the system illustrated in FIGS. 4 and 5 is provided in FIG. 8. Online processing flow corresponding to the configuration illustrated in FIG. 5 is illustrated in FIG. 8 by solid lines, while offline processing flow corresponding to the configuration illustrated in FIG. 4 is illustrated by dash-dot lines.
- Low quality source data 32 such as recorded telephone conversations, supplied by voice acquisition module(s) 50 , undergo text extraction 74 in LVCSR(s) 22 controlled by speech categorization server 60 .
- the results 76 which may include text, confidence scores, timing information and text alternative lattice information (potentially, other information, as well), undergo indexing 78 in VOIR indexing server 58 and are stored in data storage 54 .
- results 76 are supplied to speech analysis server 62 which may perform labeling 80 of the calls, as described above.
- Data from VOIR indexing server 58 are used for category training 82 , so that the categorization 84 can later be used in either online or offline mode.
- FIG. 4 One embodiment of the system illustrated in FIG. 4 is used to process recorded telephone conversations at a call center by automatically generating transcriptions of the conversations.
- offline ad-hoc querying 86 utilizes categorization 84 or rule-based keyword spotting 88 to obtain information 90 related to at least one of the documents, including at least some of the non-textual differences between n-word sequences that may be displayed on user terminal(s) 56 in the format illustrated in FIG. 8 or 9 .
- the display illustrated in FIG. 8 provides an example of user input keywords 102 “call OR meeting” that have been found in 170 documents, eight of which are displayed on screen in FIG. 8.
- the documents may be listed in a table 104 in an order based in part on the confidence of accuracy of the keywords displayed in the list.
- table 104 includes call metadata, such as start time.
- a waveform 106 of a portion of the seventh document (indicated as selected by shading in table 104 ) is displayed in the lower portion of the screen with indications of when the keywords were detected.
- the text 108 recognized by LVCSR(s) 22 .
- text 108 indicates the recognition confidence of the words and the salient terms listed in the query using one or more of highlighting, underlining, color or shade, size and style of fonts.
- labels of the conversation such as “Technical” and “Incomplete” which follow the “Categories” 116 and appear in the column under “Contact Related To” in table 104 , along with similar category information. Confidence of these labels is also indicated.
- a user may listen to the entire recording by using a pointing device, such as a computer mouse, to select a row in table 104 corresponding to the recording or can hear just the segments of audio corresponding to transcribed salient terms by selecting the speaker icon under the word “Play” on the row.
- a pointing device such as a computer mouse
- a user may select one of the words, such as “call” in a user-selectable speech bubble 110 associated with the waveform, or in the adjoining text, to skip directly to the point in a conversation where the word was said.
- a pointer 112 below the waveform 106 indicates what sound is being played back to the user and a vertical cursor 114 indicates what word was recognized for the associated sound.
- user terminal(s) 56 can also be used to graphically display results, e.g., content information 90 , indicating parameters extracted from the documents.
- results e.g., content information 90
- Examples illustrated in FIG. 9 are bar graphs 118 , 120 .
- left bar graph 118 shows the number of calls matching a query based on call date
- right graph 120 shows the relations of several categories to the user query.
- the present invention has been described with respect to embodiments using text documents generated from telephone calls.
- the invention is not limited to texts generated in this manner and can also be applied to text obtained in other ways, such as from fact extraction systems.
- the present invention can be used with any system for processing documents that derive from at least one of spontaneous and conversational expression which outputs non-deterministic text with average word recognition below 50 percent.
Abstract
Description
- This application is related to and claims priority to U.S. provisional application entitled METHOD FOR AUTOMATIC AND SEMI-AUTOMATIC CLASSIFICATION AND CLUSTERING OF NON-DETERMINISTIC TEXTS having serial No. 60/444,982, by Assaf ARIEL, Itsik HOROWITZ, Itzik STAUBER, Michael BRAND, Ofer SHOCHET and Dror ZIV, filed Feb. 5, 2003 and incorporated by reference herein. This application is also related to the application entitled AUGMENTATION AND CALIBRATION OF OUTPUT FROM NON-DETERMINISTIC TEXT GENERATORS BY MODELING ITS CHARACTERISTICS IN SPECIFIC ENVIRONMENTS by Michael BRAND, filed concurrently and incorporated by reference herein.
- 1. Field of the Invention
- The present invention is directed to processing of information in non-deterministic texts to increase the usefulness of the texts and, more particularly, to using non-textual information to indicate the importance or recognition accuracy of individual words or sequences of words.
- 2. Description of the Related Art
- In general, spoken document retrieval (SDR) is composed of two stages: transcription of speech and information retrieval (IR). Transcription of the speech is often referred to as speech-to-text (STT) or automatic speech recognition (ASR), and is often performed using a large vocabulary continuous speech recognizer (LVCSR). Information retrieval (IR) is a general term referring to all forms of data mining. One common form of data mining, for example, is query-based retrieval, where, based on a user's query, documents are retrieved and presented to the user, ordered by an estimated measure of their relevance to the query. Traditionally, this stage is performed on the text output of the first stage.
- There are many known techniques for extracting useful information from texts, commonly referred to as text mining or text data mining which is a sub-discipline of data mining. Many of these techniques have been used on text output by speech-to-text algorithms or automatic character recognition systems. However, in systems that use text that has been converted from digitized speech or is based on character recognition, there has been little success when the original source is of low quality, such as telephone conversations or handwritten text, due to the low precision of accuracy of the resulting texts. As a result, most commentators in the field have discouraged application of techniques developed for easily recognized source material to source material that is difficult to recognize. Examples of such techniques can be found in U.S. Pat. Nos. 5,625,748; 6,397,181 and 6,598,054, all incorporated by reference herein.
- Therefore, there are no known systems that provide easy access to poor quality audio, except when it is in a predictable format, such as the rules that conversations between air traffic controllers and persons in the cockpit of an aircraft follow.
- It is an aspect of the present invention to improve access to text by using non-textual information.
- It is another aspect of the present invention to use conventional text mining techniques in previously developed text mining software in a way that utilizes non-textual information in data mining.
- It is a further aspect of the present invention to improve access to documents produced by speech recognizers using recognition confidence measurement.
- The above aspects can be attained by a method for processing documents derived from at least one of spontaneous and conversational expression and containing non-deterministic text with average word recognition precision below 50 percent, the processing utilizing non-textual differences between n-word sequences in the documents to resolve more than two decision options, where n is a positive integer. Such text may be obtained by automatic character recognition or automatic speech recognition of audio signals received via a telephone system. In the preferred embodiment, the non-textual differences between the n-word sequences relate to recognition confidence of the n-word sequences
- When the processing requires fast access to the information stored in a large corpus of documents, e.g. for the purpose of data mining, the data is preferrably pre-processed to index the n-word sequences in a method that utilizes the non-textual differences between them. Such a procedure can speed up many forms of data access, and in particular many forms of data mining, including query based retrieval, as would be apparent to a person skilled in the art.
- The data mining may include extracting parameters from the documents utilizing the non-textual differences between the n-word sequences and establishing relations between the parameters extracted from the documents. The parameters extracted from the documents may be fully known, such as parameters available in document metadata or may be hidden variables that cannot be fully determined from information existing in the document. Examples of extracted parameters include an assessment of relevance to a query based on the non-textual differences between the n-word sequences and an assessment of the document's relevance to a category.
- As an alternative to creating index(es) indicating non-textual differences between n-word sequences, algorithm(s) can be used to convert text containing non-textual differences between the n-word sequences into different standard text documents. Many different algorithms may be used to transform non-deterministic text into standard text documents usable in text mining. For example, the algorithm to extract standard text documents from text with non-textual differences may apply a thresholding algorithm with varying thresholds. Then, one or more data mining techniques, each of which does not utilize non-textual differences, can be applied to these standard text documents and the outputs of the different data mining techniques can be merged to obtain information that is equivalent to that obtained by data mining that utilizes the non-textual differences.
- Whether or not the index(es) include an indication of non-textual differences, the documents may be categorized, clustered, classified, filtered or labeled, e.g., by using an algorithm to detect salient terms in the documents based on non-linguistic differences between the n-word sequences.
- In response to a query using any type(s) of index(es), information related to at least one of the documents may be displayed, including at least some non-textual differences between n-word sequences. Portions of the document(s) may be selectively displayed based on confidence of the accuracy of the displayed words. For example, salient terms in the document(s) may be displayed based on processing of confidence levels of recognition of the salient terms that resolves more than two decision options. In addition, parameters extracted from the documents and indications of the relations between these parameters may be displayed graphically.
- In response to the display of such information, a user may indicate errors in recognition. In this case at least one word in the document is preferably replaced with a corrected word supplied by the user and the confidence level(s) of the corrected word(s) are reset to indicate high recognition accuracy.
- These, together with other aspects and advantages which will be subsequently apparent, reside in the details of construction and operation as more fully hereinafter described and claimed, reference being had to the accompanying drawings forming a part hereof, wherein like numerals refer to like parts throughout.
- FIG. 1 is a flowchart of a conventional spoken document retrieval system.
- FIG. 2 is a flowchart of one method of spoken document retrieval according to the invention.
- FIG. 3 is a flowchart of another method of spoken document retrieval according to the invention.
- FIGS. 4 and 5 are block diagrams of spoken document retrieval systems according to the invention.
- FIG. 6 is a block diagram of one confidence sensitive inverted index and one regular index containing confidence information.
- FIG. 7 is a flowchart of text processing according to the invention.
- FIGS. 8 and 9 are examples of displays generated by telephone call processing applications according to the invention.
- Following are several terms used herein that are in common use in automatic speech recognition or data mining.
labeling a form of data processing where documents are analyzed, and the analysis results (referred to as “labels”) are made available for later processing stages. For example, a topical analysis of documents is a labeling of the documents by subject. retrieving a form of data mining where a subset of a document corpus is returned in response to a query. Preferably, the documents are each given a rank pertaining to their relevance to the query, and are sorted by decreasing relevance. categorizing a form of data mining where several “categories” are defined, and the documents of a corpus are labeled according the category to which they fit best. A common variation is multilabel categorizing, where each document may fit zero or more categories. Preferably, information is given regarding the quality of the fit. clustering a form of data mining similar to categorization, with the difference that the “categories” are not predefined, and the data mining must reveal them automatically. classifying a process performed on a stream of incoming documents, where each is labeled and then forwarded for relevant additional processing (manual or automatic) based on the labels that have been discovered. filtering a process performed on a stream of incoming documents, where each is labeled and then forwarded or discarded based on the labels that have been discovered. salient terms terms whose appearance in a document provides information relevant to its correct labeling, and consequently to all forms of data mining subsequent to labeling. - First, processing performed by a typical spoken document retrieval system will be described with reference to FIG. 1.
High quality audio 20 is input into anASR system 22 using LVCSR.ASR system 22 converts spoken words into textually represented words, but often has other outputs as well. These outputs may include timing information, an indication of the confidence of recognition of particular words and phrases, alternative likely transcriptions, and more. - The LVCSR output cannot be piped directly into a traditional text mining system. It has to be converted into searchable text. For this reason, canonization24 is performed to produce canonized text, also referred to below as standard text documents, used by conventional text mining software. Most commonly, canonization simply involves taking the textual words out of the LVCSR output and concatenating them. More sophisticated canonization schemes involve usage of both textual and non-textual information to convert the LVCSR output into a format more easily handled by
text mining system 26. Usage of textual information may include capitalization and punctuation based on grammatical rules. Usage of non-textual information may include capitalization and punctuation using timing information and word(s) omitted based on low confidence levels. -
Text mining system 26 receives input from many different audio segments and stores the information in some format that will be convenient for later processing, a process known as “indexing”. When asked to produce output, typically, though not exclusively, byuser query 28,text mining system 26 searches its index and produces output. For example the output may be the identities of the audio segments that were requested for retrieval, ranked and scored by some relevance metric. The output may also include other information, such as the phrases in the retrieved segments that have proved to be salient terms, because of which the document was given the score that it was given. - All this information is finally piped into
document display system 30 which can use all of it, and add to it the original audio segment(s) 20, to give the user feedback touser query 28 that is as informative and audio-visually appealing as possible. - A simplified embodiment of a method according to the present invention for spoken document retrieval of
low grade audio 32 is illustrated in FIG. 2. A similar system could be used for text generated by a handwritten text recognition system. In embodiment illustrated in FIG. 2, the traditional text mining system has been replaced withdata mining system 34 that is designed especially for speech inlow grade audio 32.Data mining system 34 doesn't require that the output ofASR 36 be canonized into text before it is handled, and can therefore utilize all information available in the output ofASR 36. No information is lost in a canonization process, and words that are indexed can receive different and appropriate handling based on non-textual information, such as their confidence scores. Thecanonization stage 24 of the process illustrated in FIG. 1 is therefore entirely omitted, and the output ofASR 36 is available with more of its information in subsequent processing stages, including indexing, retrieval and display. Since non-textual information is available,document display system 38 in FIG. 2 displays different information thandocument display system 30 in FIG. 1. - Traditionally, speech data mining was confined to
high quality audio 20 such as broadcast quality audio. Broadcast audio is of high quality, typically achieving 60-80% word recognition precision, and certainly better than 50%. By contrast, telephony grade data typically has 20-30% word recognition precision, and certainly worse than 50%. Traditional SDR systems do not utilize non-linguistic information during data mining (in all stages where documents are handled together and not separately). Attempts to use traditional text mining methodologies on low-grade audio data have proven inadequate, with the results not good enough for commercial use. -
Data mining system 34 uses data mining targeted specifically at low-grade audio speech data, extracting useful information by ranking and retrieval utilizing non-textual data which makes data mining of telephony speech quite feasible. In the embodiment illustrated in FIG. 2,data mining 34 includes indexing which differentiates words using non-textual differences between the words. Upon receipt ofuser query 28,data mining system 34 uses the non-textual differences between the words for retrieval. - An alternative way to determine non-textual differences between words without using a specially created data mining system is illustrated in FIG. 3. In this scheme, the same segments of
audio 32 are canonized several times usingdifferent canonization methods 24 a, . . . 24 n. These methods can differ, for example, by a choice of minimal allowed confidence, any word of confidence lower than that being removed from the transcription in one way or another. - After applying the N
different canonization schemes 24 a, . . . 24 n, the obtained text is indexed in N different indexes bytext mining systems 26 a, . . . 26 n which may use the same indexing methods or different methods. Although not illustrated in FIG. 3,low grade audio 32 may additionally be processed byASR 36 to produce output for data mining system(s) 34 that utilize(s) non-textual differences between words. - When user query28 is received,
text mining systems 26 a, . . . 26 n search indexes of standard text documents, result(s) are retrieved from each index, and finally all N results are algorithmically merged 42, providing a single output, which is then forwarded to documentdisplay system 38. The merged output, unlike the single outputs of the text mining systems, can differentiate results by confidence levels, anddisplay system 38 can use this information. - Either of the spoken document retrieval methods illustrated in FIGS. 2 and 3 can be implemented using systems like that illustrated in FIGS. 4 and 5. Such systems can be configured in many different ways, depending on the tasks needed to be performed and the volume of data processed and FIGS. 4 and 5 are just two examples. The invention is not limited to the configurations illustrated in FIGS. 4 and 5 and other configurations are possible and will be apparent to a person of ordinary skill in the art. For example, the functions performed by separate servers in FIGS. 4 and 5 may be performed by separate modules in a single computing system.
- In the configuration illustrated in FIG. 4, a system according to the invention is used interactively offline. Voice data are supplied from voice acquisition module(s)50 by
network 52 and stored indata storage 54.Network 52 may be any known type of network, such as a local area network (LAN), wide area network (WAN), the Internet, etc. The voice data may be in the form of WAV files or any other audio file format. - In either of the configurations illustrated in FIGS. 4 and 5, the system is accessed by one or
more user terminals 56, such as personal computers or other devices that include a user interface which may include a display. In the interactive offline system illustrated in FIG. 4, users log into the system at various times to submit queries to voice oriented information retrieval (VOIR) indexingserver 58. Voice data from voice acquisition module(s) 50 are supplied tospeech categorization server 60 which, if necessary, converts the data before supplying the voice data to LVCSR(s) 22 and performs load balancing when more than oneLVCSR 22 is used. - LVCSR(s)22 output words and additional data, such as speaker-change, timing information, confidence scores, etc. In addition, call metadata, such as the time that a call was made and the number dialed, is obtained from voice acquisition module(s) 50 together with the voice data. All these types of data are combined, e.g., by
speech categorization server 60 and forwarded, in online mode tospeech analysis server 62 and in offline mode toVOIR indexing server 58. Regardless of whether the method illustrated in FIG. 2 or FIG. 3 is implemented, results of a query in offline mode can be displayed onuser terminals 56 with at least some of the non-textual differences between n-word sequences indicated. Examples of how the non-textual differences are conveyed to the user will be described below with reference to FIGS. 7 and 8. - The online configuration illustrated in FIG. 5, may be used when the volume of voice data is too large to allow effective offline processing, or it is desired to use push-technology alerts to people who may want the data. For example, a police inspector may want to be paged when the system detects a phone conversation relevant to her case. In the online configuration illustrated in FIG. 5, the output of LVCSR(s)22 is supplied via
network 52 tospeech analysis server 62 which labels the voice data. For example, the voice data may be labeled according to importance, subject matter, person or group that needs to respond, etc. The labeling of the transcribed voice data is combined with the output ofLVCSRs 22 and call metadata, and forwarded to categorization queue andworkflow manager 64. The users atuser terminals 56 are provided this information by categorization queue andworkflow manager 64. Using the labeling provided byspeech analysis server 62, categorization queue andworkflow manager 64 supplies text, voice data and call metadata appropriate for that user, depending on importance, topic, identity of the user, etc. - Training of
speech analysis server 62 may be accomplished by offline processing usingVOIR indexing server 58 in an implementation that includes bothservers speech analysis server 62 as training data to recognize similar calls during online processing of calls in a call center, for example. In addition, training may continue during online processing as users correct the labeling provided byspeech analysis server 62. When all processing is offline,VOIR indexing server 58 is trained in a similar manner. - In a typical implementation of the invention, a
single LVCSR 22 pass is sufficient for each call. If the method described above with reference to FIG. 2 is implemented,LVCSR 22 supplies metadata, including confidence scores, associated with recognized words toVOIR indexing server 58 which generates an index that indicates at least some of the non-textual differences between n-word sequences. If the method described above with reference to FIG. 3 is implemented,VOIR indexing server 58 maintains an index for eachcanonization system 24. In either case, the index(es) and the voice data (preferably compressed to minimize space requirements) or other data from which indexed text is obtained (such as handwritten documents) are preferably stored indata storage 54. - In the preferred embodiment, if the method illustrated in FIG. 2 is implemented, data in
data storage 54 is indexed by use of at least one confidence sensitive inverted index. A confidence sensitive inverted index maps from terms to a sorted linked list identifying all documents where each term occurs and from each appearance of a document in this list to a sorted linked list identifying all positions in which the term appears and the confidence level of its recognition. In addition (or alternatively), indexed data may include aggregated information relating to confidence. - An example is illustrated in FIG. 6 of a confidence sensitive
inverted index 65 and a regular (forward)index 66 containing confidence information with the twoindexes mapping 67 i) in the indexed document (e.g., 68 a) can be calculated based on the individual recognition scores of the occurrences. - Another example of aggregated information relating to confidence information that can be saved is the strength of association between every document and each category. This information can be saved either in a regular (forward) index, like
index 66, another inverted index (not shown), or both. Information not relating to confidence, such as call metadata, can also be indexed, either in another inverted index, in a forward index likeindex 66, or both. In both cases, if an inverted index is used, confidence sensitiveinverted index 65 or a separate index can be used. Furthermore, additional mapping technologies, in addition to or instead of a mapping into a sorted linked list, can also be used.Data storage 54 can also store information other than the indexes, such as the data that is being indexed. This data may include, among others, call audio, voice data and call metadata, and may include additional indexes used to refer to the same data. - A more detailed flow of processing through the system illustrated in FIGS. 4 and 5 is provided in FIG. 8. Online processing flow corresponding to the configuration illustrated in FIG. 5 is illustrated in FIG. 8 by solid lines, while offline processing flow corresponding to the configuration illustrated in FIG. 4 is illustrated by dash-dot lines. Low
quality source data 32, such as recorded telephone conversations, supplied by voice acquisition module(s) 50, undergotext extraction 74 in LVCSR(s) 22 controlled byspeech categorization server 60. In offline mode, theresults 76, which may include text, confidence scores, timing information and text alternative lattice information (potentially, other information, as well), undergoindexing 78 inVOIR indexing server 58 and are stored indata storage 54. In the online mode, results 76 are supplied tospeech analysis server 62 which may perform labeling 80 of the calls, as described above. Data fromVOIR indexing server 58 are used forcategory training 82, so that thecategorization 84 can later be used in either online or offline mode. - One embodiment of the system illustrated in FIG. 4 is used to process recorded telephone conversations at a call center by automatically generating transcriptions of the conversations. In this embodiment, offline ad-hoc querying86 (FIG. 8) utilizes
categorization 84 or rule-based keyword spotting 88 to obtaininformation 90 related to at least one of the documents, including at least some of the non-textual differences between n-word sequences that may be displayed on user terminal(s) 56 in the format illustrated in FIG. 8 or 9. The display illustrated in FIG. 8 provides an example ofuser input keywords 102 “call OR meeting” that have been found in 170 documents, eight of which are displayed on screen in FIG. 8. Preferably, the documents may be listed in a table 104 in an order based in part on the confidence of accuracy of the keywords displayed in the list. In the example illustrated in FIGS. 8 and 9, table 104 includes call metadata, such as start time. - In the example illustrated in FIG. 8, a
waveform 106 of a portion of the seventh document (indicated as selected by shading in table 104) is displayed in the lower portion of the screen with indications of when the keywords were detected. Below the waveform is thetext 108 recognized by LVCSR(s) 22. Preferably,text 108 indicates the recognition confidence of the words and the salient terms listed in the query using one or more of highlighting, underlining, color or shade, size and style of fonts. Also shown in the example illustrated in FIGS. 8 and 9 are labels of the conversation, such as “Technical” and “Incomplete” which follow the “Categories” 116 and appear in the column under “Contact Related To” in table 104, along with similar category information. Confidence of these labels is also indicated. - In one embodiment of the invention, a user may listen to the entire recording by using a pointing device, such as a computer mouse, to select a row in table104 corresponding to the recording or can hear just the segments of audio corresponding to transcribed salient terms by selecting the speaker icon under the word “Play” on the row. Once a row has been selected, a user may select one of the words, such as “call” in a user-
selectable speech bubble 110 associated with the waveform, or in the adjoining text, to skip directly to the point in a conversation where the word was said. Apointer 112 below thewaveform 106 indicates what sound is being played back to the user and avertical cursor 114 indicates what word was recognized for the associated sound. - Preferably, user terminal(s)56 can also be used to graphically display results, e.g.,
content information 90, indicating parameters extracted from the documents. Examples illustrated in FIG. 9 arebar graphs bar graph 118 shows the number of calls matching a query based on call date, whileright graph 120 shows the relations of several categories to the user query. - The present invention has been described with respect to embodiments using text documents generated from telephone calls. However, as noted above, the invention is not limited to texts generated in this manner and can also be applied to text obtained in other ways, such as from fact extraction systems. Furthermore, the present invention can be used with any system for processing documents that derive from at least one of spontaneous and conversational expression which outputs non-deterministic text with average word recognition below 50 percent.
- The many features and advantages of the invention are apparent from the detailed specification and, thus, it is intended by the appended claims to cover all such features and advantages of the invention that fall within the true spirit and scope of the invention. Further, since numerous modifications and changes will readily occur to those skilled in the art, it is not desired to limit the invention to the exact construction and operation illustrated and described, and accordingly all suitable modifications and equivalents may be resorted to, falling within the scope of the invention.
Claims (105)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/771,315 US20040163035A1 (en) | 2003-02-05 | 2004-02-05 | Method for automatic and semi-automatic classification and clustering of non-deterministic texts |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US44498203P | 2003-02-05 | 2003-02-05 | |
US10/771,315 US20040163035A1 (en) | 2003-02-05 | 2004-02-05 | Method for automatic and semi-automatic classification and clustering of non-deterministic texts |
Publications (1)
Publication Number | Publication Date |
---|---|
US20040163035A1 true US20040163035A1 (en) | 2004-08-19 |
Family
ID=32869303
Family Applications (4)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/771,315 Abandoned US20040163035A1 (en) | 2003-02-05 | 2004-02-05 | Method for automatic and semi-automatic classification and clustering of non-deterministic texts |
US10/771,409 Active 2027-03-11 US7792671B2 (en) | 2003-02-05 | 2004-02-05 | Augmentation and calibration of output from non-deterministic text generators by modeling its characteristics in specific environments |
US12/059,660 Abandoned US20080183468A1 (en) | 2003-02-05 | 2008-03-31 | Augmentation and calibration of output from non-deterministic text generators by modeling its characteristics in specific environments |
US12/876,207 Expired - Lifetime US8195459B1 (en) | 2003-02-05 | 2010-09-06 | Augmentation and calibration of output from non-deterministic text generators by modeling its characteristics in specific environments |
Family Applications After (3)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/771,409 Active 2027-03-11 US7792671B2 (en) | 2003-02-05 | 2004-02-05 | Augmentation and calibration of output from non-deterministic text generators by modeling its characteristics in specific environments |
US12/059,660 Abandoned US20080183468A1 (en) | 2003-02-05 | 2008-03-31 | Augmentation and calibration of output from non-deterministic text generators by modeling its characteristics in specific environments |
US12/876,207 Expired - Lifetime US8195459B1 (en) | 2003-02-05 | 2010-09-06 | Augmentation and calibration of output from non-deterministic text generators by modeling its characteristics in specific environments |
Country Status (4)
Country | Link |
---|---|
US (4) | US20040163035A1 (en) |
EP (2) | EP1590798A2 (en) |
IL (1) | IL170065A (en) |
WO (2) | WO2004072780A2 (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060067578A1 (en) * | 2004-09-30 | 2006-03-30 | Fuji Xerox Co., Ltd. | Slide contents processor, slide contents processing method, and storage medium storing program |
US20070050445A1 (en) * | 2005-08-31 | 2007-03-01 | Hugh Hyndman | Internet content analysis |
US20080027888A1 (en) * | 2006-07-31 | 2008-01-31 | Microsoft Corporation | Optimization of fact extraction using a multi-stage approach |
US20090012970A1 (en) * | 2007-07-02 | 2009-01-08 | Dror Daniel Ziv | Root cause analysis using interactive data categorization |
US20090248647A1 (en) * | 2008-03-25 | 2009-10-01 | Omer Ziv | System and method for the quality assessment of queries |
US20090249253A1 (en) * | 2008-03-31 | 2009-10-01 | Palm, Inc. | Displaying mnemonic abbreviations for commands |
US20110072052A1 (en) * | 2008-05-28 | 2011-03-24 | Aptima Inc. | Systems and methods for analyzing entity profiles |
US8725732B1 (en) * | 2009-03-13 | 2014-05-13 | Google Inc. | Classifying text into hierarchical categories |
US20210312123A1 (en) * | 2020-04-03 | 2021-10-07 | Jon Ward | Systems and Methods For Cloud-Based Productivity Tools |
US11244011B2 (en) * | 2015-10-23 | 2022-02-08 | International Business Machines Corporation | Ingestion planning for complex tables |
US20230028717A1 (en) * | 2020-08-27 | 2023-01-26 | Capital One Services, Llc | Representing Confidence in Natural Language Processing |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1590798A2 (en) * | 2003-02-05 | 2005-11-02 | Verint Systems Inc. | Method for automatic and semi-automatic classification and clustering of non-deterministic texts |
US7856355B2 (en) * | 2005-07-05 | 2010-12-21 | Alcatel-Lucent Usa Inc. | Speech quality assessment method and system |
US20070078806A1 (en) * | 2005-10-05 | 2007-04-05 | Hinickle Judith A | Method and apparatus for evaluating the accuracy of transcribed documents and other documents |
US20100027768A1 (en) * | 2006-11-03 | 2010-02-04 | Foskett James J | Aviation text and voice communication system |
US8126891B2 (en) * | 2008-10-21 | 2012-02-28 | Microsoft Corporation | Future data event prediction using a generative model |
US8379801B2 (en) * | 2009-11-24 | 2013-02-19 | Sorenson Communications, Inc. | Methods and systems related to text caption error correction |
US9070360B2 (en) * | 2009-12-10 | 2015-06-30 | Microsoft Technology Licensing, Llc | Confidence calibration in automatic speech recognition systems |
US8930189B2 (en) | 2011-10-28 | 2015-01-06 | Microsoft Corporation | Distributed user input to text generated by a speech to text transcription service |
US9870520B1 (en) * | 2013-08-02 | 2018-01-16 | Intuit Inc. | Iterative process for optimizing optical character recognition |
FR3010809B1 (en) * | 2013-09-18 | 2017-05-19 | Airbus Operations Sas | METHOD AND DEVICE FOR AUTOMATIC MANAGEMENT ON BOARD AN AIRCRAFT AUDIO MESSAGE AIRCRAFT. |
US11481087B2 (en) * | 2014-03-27 | 2022-10-25 | Sony Corporation | Electronic device and method for identifying input commands of a user |
US9858923B2 (en) * | 2015-09-24 | 2018-01-02 | Intel Corporation | Dynamic adaptation of language models and semantic tracking for automatic speech recognition |
CN108777141B (en) * | 2018-05-31 | 2022-01-25 | 康键信息技术(深圳)有限公司 | Test apparatus, test method, and storage medium |
CN110110303A (en) * | 2019-03-28 | 2019-08-09 | 苏州八叉树智能科技有限公司 | Newsletter archive generation method, device, electronic equipment and computer-readable medium |
CN111581455B (en) * | 2020-04-28 | 2023-03-21 | 北京字节跳动网络技术有限公司 | Text generation model generation method and device and electronic equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5625748A (en) * | 1994-04-18 | 1997-04-29 | Bbn Corporation | Topic discriminator using posterior probability or confidence scores |
US6397181B1 (en) * | 1999-01-27 | 2002-05-28 | Kent Ridge Digital Labs | Method and apparatus for voice annotation and retrieval of multimedia data |
US20020178002A1 (en) * | 2001-05-24 | 2002-11-28 | International Business Machines Corporation | System and method for searching, analyzing and displaying text transcripts of speech after imperfect speech recognition |
US6598054B2 (en) * | 1999-01-26 | 2003-07-22 | Xerox Corporation | System and method for clustering data objects in a collection |
US20040083101A1 (en) * | 2002-10-23 | 2004-04-29 | International Business Machines Corporation | System and method for data mining of contextual conversations |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5550930A (en) * | 1991-06-17 | 1996-08-27 | Microsoft Corporation | Method and system for training a handwriting recognizer at the time of misrecognition |
GB9709341D0 (en) * | 1997-05-08 | 1997-06-25 | British Broadcasting Corp | Method of and apparatus for editing audio or audio-visual recordings |
WO2001074042A2 (en) | 2000-03-24 | 2001-10-04 | Dragon Systems, Inc. | Lexical analysis of telephone conversations with call center agents |
US6839667B2 (en) * | 2001-05-16 | 2005-01-04 | International Business Machines Corporation | Method of speech recognition by presenting N-best word candidates |
US6963834B2 (en) * | 2001-05-29 | 2005-11-08 | International Business Machines Corporation | Method of speech recognition using empirically determined word candidates |
EP1590798A2 (en) * | 2003-02-05 | 2005-11-02 | Verint Systems Inc. | Method for automatic and semi-automatic classification and clustering of non-deterministic texts |
-
2004
- 2004-01-05 EP EP04700238A patent/EP1590798A2/en not_active Ceased
- 2004-01-05 WO PCT/US2004/003198 patent/WO2004072780A2/en active Application Filing
- 2004-02-05 WO PCT/US2004/003199 patent/WO2004072955A1/en active Application Filing
- 2004-02-05 EP EP04708538A patent/EP1590796A1/en not_active Ceased
- 2004-02-05 US US10/771,315 patent/US20040163035A1/en not_active Abandoned
- 2004-02-05 US US10/771,409 patent/US7792671B2/en active Active
-
2005
- 2005-08-03 IL IL170065A patent/IL170065A/en active IP Right Grant
-
2008
- 2008-03-31 US US12/059,660 patent/US20080183468A1/en not_active Abandoned
-
2010
- 2010-09-06 US US12/876,207 patent/US8195459B1/en not_active Expired - Lifetime
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5625748A (en) * | 1994-04-18 | 1997-04-29 | Bbn Corporation | Topic discriminator using posterior probability or confidence scores |
US6598054B2 (en) * | 1999-01-26 | 2003-07-22 | Xerox Corporation | System and method for clustering data objects in a collection |
US6397181B1 (en) * | 1999-01-27 | 2002-05-28 | Kent Ridge Digital Labs | Method and apparatus for voice annotation and retrieval of multimedia data |
US20020178002A1 (en) * | 2001-05-24 | 2002-11-28 | International Business Machines Corporation | System and method for searching, analyzing and displaying text transcripts of speech after imperfect speech recognition |
US20040083101A1 (en) * | 2002-10-23 | 2004-04-29 | International Business Machines Corporation | System and method for data mining of contextual conversations |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060067578A1 (en) * | 2004-09-30 | 2006-03-30 | Fuji Xerox Co., Ltd. | Slide contents processor, slide contents processing method, and storage medium storing program |
US7698645B2 (en) * | 2004-09-30 | 2010-04-13 | Fuji Xerox Co., Ltd. | Presentation slide contents processor for categorizing presentation slides and method for processing and categorizing slide contents |
US20070050445A1 (en) * | 2005-08-31 | 2007-03-01 | Hugh Hyndman | Internet content analysis |
US20080027888A1 (en) * | 2006-07-31 | 2008-01-31 | Microsoft Corporation | Optimization of fact extraction using a multi-stage approach |
US7668791B2 (en) | 2006-07-31 | 2010-02-23 | Microsoft Corporation | Distinguishing facts from opinions using a multi-stage approach |
US9015194B2 (en) * | 2007-07-02 | 2015-04-21 | Verint Systems Inc. | Root cause analysis using interactive data categorization |
US20090012970A1 (en) * | 2007-07-02 | 2009-01-08 | Dror Daniel Ziv | Root cause analysis using interactive data categorization |
US20090248647A1 (en) * | 2008-03-25 | 2009-10-01 | Omer Ziv | System and method for the quality assessment of queries |
US20090249253A1 (en) * | 2008-03-31 | 2009-10-01 | Palm, Inc. | Displaying mnemonic abbreviations for commands |
US9053088B2 (en) * | 2008-03-31 | 2015-06-09 | Qualcomm Incorporated | Displaying mnemonic abbreviations for commands |
US20110072052A1 (en) * | 2008-05-28 | 2011-03-24 | Aptima Inc. | Systems and methods for analyzing entity profiles |
US9123022B2 (en) | 2008-05-28 | 2015-09-01 | Aptima, Inc. | Systems and methods for analyzing entity profiles |
US9594825B2 (en) | 2008-05-28 | 2017-03-14 | Aptima, Inc. | Systems and methods for analyzing entity profiles |
US11461373B2 (en) | 2008-05-28 | 2022-10-04 | Aptima, Inc. | Systems and methods for analyzing entity profiles |
US8725732B1 (en) * | 2009-03-13 | 2014-05-13 | Google Inc. | Classifying text into hierarchical categories |
US11244011B2 (en) * | 2015-10-23 | 2022-02-08 | International Business Machines Corporation | Ingestion planning for complex tables |
US20210312123A1 (en) * | 2020-04-03 | 2021-10-07 | Jon Ward | Systems and Methods For Cloud-Based Productivity Tools |
US11687710B2 (en) * | 2020-04-03 | 2023-06-27 | Braincat, Inc. | Systems and methods for cloud-based productivity tools |
US20230028717A1 (en) * | 2020-08-27 | 2023-01-26 | Capital One Services, Llc | Representing Confidence in Natural Language Processing |
US11720753B2 (en) * | 2020-08-27 | 2023-08-08 | Capital One Services, Llc | Representing confidence in natural language processing |
Also Published As
Publication number | Publication date |
---|---|
WO2004072955A1 (en) | 2004-08-26 |
EP1590798A2 (en) | 2005-11-02 |
IL170065A (en) | 2013-02-28 |
US20080183468A1 (en) | 2008-07-31 |
US8195459B1 (en) | 2012-06-05 |
US20040158469A1 (en) | 2004-08-12 |
US7792671B2 (en) | 2010-09-07 |
WO2004072780A3 (en) | 2004-11-11 |
WO2004072780A2 (en) | 2004-08-26 |
EP1590796A1 (en) | 2005-11-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20040163035A1 (en) | Method for automatic and semi-automatic classification and clustering of non-deterministic texts | |
US10431214B2 (en) | System and method of determining a domain and/or an action related to a natural language input | |
CN108197282B (en) | File data classification method and device, terminal, server and storage medium | |
US7272558B1 (en) | Speech recognition training method for audio and video file indexing on a search engine | |
US8239189B2 (en) | Method and system for estimating a sentiment for an entity | |
US7415409B2 (en) | Method to train the language model of a speech recognition system to convert and index voicemails on a search engine | |
US20190278812A1 (en) | Model generation device, text search device, model generation method, text search method, data structure, and program | |
CN111046133A (en) | Question-answering method, question-answering equipment, storage medium and device based on atlas knowledge base | |
US20100070263A1 (en) | Speech data retrieving web site system | |
US20040249808A1 (en) | Query expansion using query logs | |
CN105956053B (en) | A kind of searching method and device based on the network information | |
CN107748784B (en) | Method for realizing structured data search through natural language | |
JP2013521567A (en) | System including client computing device, method of tagging media objects, and method of searching a digital database including audio tagged media objects | |
CN112800170A (en) | Question matching method and device and question reply method and device | |
CN109446376B (en) | Method and system for classifying voice through word segmentation | |
CN108038099B (en) | Low-frequency keyword identification method based on word clustering | |
CN102081634A (en) | Speech retrieval device and method | |
US20220058213A1 (en) | Systems and methods for identifying dynamic types in voice queries | |
WO2006118360A1 (en) | Issue trend analysis system | |
CN111881283A (en) | Business keyword library creating method, intelligent chat guiding method and device | |
CN111125457A (en) | Deep cross-modal Hash retrieval method and device | |
US20230214579A1 (en) | Intelligent character correction and search in documents | |
CN111488453B (en) | Resource grading method, device, equipment and storage medium | |
JP6576847B2 (en) | Analysis system, analysis method, and analysis program | |
CN111090977A (en) | Intelligent writing system and intelligent writing method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: VERINT SYSTEMS INC., NEW YORK Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ARIEL, ASSAF;BRAND, MICHAEL;HOROWITZ, ITSIK;AND OTHERS;REEL/FRAME:014967/0368;SIGNING DATES FROM 20040129 TO 20040202 |
|
AS | Assignment |
Owner name: LEHMAN COMMERCIAL PAPER INC., AS ADMINISTRATIVE AG Free format text: SECURITY AGREEMENT;ASSIGNOR:VERINT SYSTEMS INC.;REEL/FRAME:019588/0613 Effective date: 20070525 |
|
AS | Assignment |
Owner name: CREDIT SUISSE AS ADMINISTRATIVE AGENT, NEW YORK Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:VERINT SYSTEMS INC.;LEHMAN COMMERCIAL PAPER INC.;REEL/FRAME:022793/0888 Effective date: 20090604 |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |
|
AS | Assignment |
Owner name: VERINT AMERICAS INC., NEW YORK Free format text: RELEASE OF SECURITY INTEREST;ASSIGNOR:CREDIT SUISSE AG;REEL/FRAME:026206/0340 Effective date: 20110429 Owner name: VERINT SYSTEMS INC., NEW YORK Free format text: RELEASE OF SECURITY INTEREST;ASSIGNOR:CREDIT SUISSE AG;REEL/FRAME:026206/0340 Effective date: 20110429 Owner name: VERINT VIDEO SOLUTIONS INC., NEW YORK Free format text: RELEASE OF SECURITY INTEREST;ASSIGNOR:CREDIT SUISSE AG;REEL/FRAME:026206/0340 Effective date: 20110429 |