International Business Machines Corporation

Noam Slonim

Ehud Aharoni

Ruty Rinott

Yonatan Bilu

Dan Gutfreund

Ran Levy

Ranit Aharonov

Roy N Bar-Haim

Ilya Shnayderman

Tamar Lavee

Daniel Hershcovich

Orith Toledo-Ronen

Lena Dankin

Liat Ein Dor

Alon Halfon

Oded Margalit

Robert J Baseman

Indrajit Bhattacharya

Amrita Saha

Yosi Mass

Artem Spector

Vikas Raykar

Raminderpal Singh

Adam D Ticknor

Hong Lin

Fateh A Tipu

Michal Rosen-Zvi

Ramona Kei

Eyal Shnarch

Timothy M McCormack

Kevin MacKey

Amir Menczel

Ashish Verma

David Carmel

Hui Zhu

Elad Yom-Tov

David Konopnicki

Feng Cao

Boaz Carmeli

Yuan Ni

Charles Arthur Jochim

Dafna Sheinwald

Ella Rabinovich

Benjamin Sznajder

Chen Yanover

Qiongkai Xu

Yoav Kantor

Mitesh Khapra

Lilach Eden

Liat Ein-Dor

Godbole Shantanu

Francesco Dinuzzo

Aharoni Ehud

Assaf Toledo

Edo Cohen-Karlik

Avishai Gretz

Dalia Krieger

Ariel Gera

Leshem Choshen

Elad Venezian

An example system includes a processor to receive a list of sentiment carrying discourse markers. The processor is to select sentences in a text corpus that begin with a discourse marker from the list of sentiment carrying discourse markers followed by a comma. The processor is to remove each discourse marker and comma from a beginning of the selected sentences and labeling each of the sentences with a sentiment associated with to a corresponding removed discourse marker to generate a weakly labeled dataset. The processor is to inter-train a pretrained language model using the generated weakly labeled dataset to generate a sentiment model.

Generating sentiment models using weak labels generated from discourse markers

An example system includes a processor to pre-train a transformer-based language model on a general domain. The processor can inter-train the pre-trained transformer-based language model using partitioning and classification to generate an inter-trained transformer-based pre-trained language model. The processor can then fine-tune the inter-trained transformer-based pre-trained language model on a target task to generate a fine-tuned transformer-based language model.

Inter-training of pre-trained transformer-based language models using partitioning and classification

A computing system, computer program product, and computer-implemented method for quantitative comment summarization are provided. The method includes receiving a collection of comments, identifying a set of candidate key points corresponding to the collection of comments, and selecting a subset of key points from the set of candidate key points, wherein the selected subset of key points includes key points that are most salient in the collection of comments. The method also includes automatically mapping each comment within the collection of comments to any corresponding key points within the subset of key points based on a match score between each comment and key point pair, as well as generating a summary including the subset of key points and an absolute number or percentage of the comments mapped to each key point.

Quantitative comment summarization

Automated detection of reasoning in arguments. A training set is generated by: obtaining multiple arguments, each comprising one or more sentences provided as digital text; automatically estimating a probability that each of the arguments includes reasoning, wherein the estimating comprises applying a contextual language model to each of the arguments; automatically labeling as positive examples those of the arguments which have a relatively high probability to include reasoning; and automatically labeling as negative examples those of the arguments which have a relatively low probability to include reasoning. Based on the generated training set, a machine learning classifier is automatically trained to estimate a probability that a new argument includes reasoning. The trained machine learning classifier is applied to the new argument, to estimate a probability that the new argument includes reasoning.

Automated detection of reasoning in arguments

Embodiments may provide techniques for clustering using predefined anchors that take into account the knowledge about the anchors. For example, a method of clustering processing may be implemented in a computer comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the method comprising: receiving a plurality of points of data to be clustered and a plurality of predefined anchor data points and clustering the plurality of points of data and at least a subset of the predefined anchor data points. Further, the method may include creating a similarity function where anchor points pull stronger than regular points, such that they function as attractors.

Clustering items around predefined anchors

A computerized text analysis method that comprises: searching a resource of information with a search query comprising at least one of: (a) the specific debatable topic, and (b) a personal derivation of the specific debatable topic, to obtain a list of indices whose index subject contains the personal derivation and/or the specific debatable topic; determining, by applying a rule-based classifier, whether the index subject of each of the indices is (i) in favor of the debatable topic or (ii) against the debatable topic; detecting, in each of the indices, hyperlinks to encyclopedic entries whose entry subjects are person names; and determining that: if the index subject of each of the one or more indices is in favor of the specific debatable topic, then the persons are in favor of the specific debatable topic, and vice versa.

Expert stance classification using computerized text analytics

A method including: generating, from a text corpus, a lexicon of unigrams and bigrams comprising an embedding for each of said unigrams and bigrams; training a machine learning classifier on a training set comprising a subset of said lexicon, wherein each of said unigrams and bigrams in said subset has a sentiment label; applying said machine learning classifier to said lexicon, to (i) predict a sentiment of each of said unigrams and bigrams, and (ii) update said lexicon with the predicted sentiments; and performing statistical analysis on said updated lexicon, to extract one or more sentiment composition lexicons, wherein each of said one or more sentiment composition lexicons is associated with a sentiment composition class.

Learning sentiment composition from sentiment lexicons

A system for identifying in a corpus of documents at least one argument relevant to an identified topic, comprising at least one hardware processor adapted to: producing a plurality of topic-related sentences relevant to the identified topic, each extracted from a document of the corpus of documents; producing a plurality of synthetic documents, each created by appending to a sentence of the plurality of topic-related sentences an identified amount of other sentences extracted from the respective document the topic-related sentence was extracted therefrom; identifying at least one argument relevant to the identified topic by inputting each of the plurality of synthetic documents to at least one machine learning model trained to identify an argument in response to a document; and outputting the at least one argument.

System and method for argument retrieval

A method comprising using at least one hardware processor for: receiving a topic under consideration (TUC) and content relevant to the TUC; detecting one or more claims relevant to the TUC in the content, based on detection of boundaries of the claims in the content; and outputting a list of said detected one or more claims.

Automatic detection of claims with respect to a topic

Method and apparatus for training and using a classifier for words. Embodiments include receiving a first plurality of sentences comprising a first word that is associated with a class and a second plurality of sentences comprising a second word that is not associated with the class. Embodiments include training a classifier using positive training data for the class that is based on the first plurality of sentences and negative training data for the class that is based on the second plurality of sentences. Embodiments include determining a measure of correlation between a third word and the class by using a sentence comprising the third word as an input to the classifier. Embodiments include using the measure of correlation to perform an action selected from the following list: selecting content to provide to a user; determining an automatic chat response; or filtering a set of content.

Classification engine for learning properties of words and multi-word expressions

A method of estimating a thematic similarity of sentences, comprising receiving a corpus of a plurality of documents describing a plurality of topics where each document comprises a plurality of sentences arranged in a plurality of sections, constructing sentence triplets for at least some of the sentences, each sentence triplet comprising a respective sentence, a respective positive sentence selected randomly from the section comprising the respective sentence and a respective negative sentence selected randomly from another section, training a first neural network with the sentence triplets to identify sentence-sentence vectors mapping each sentence with a shorter distance to its respective positive sentence compared to the distance to its respective negative sentence and outputting the first neural network for estimating thematic similarity between a pair of sentences by computing a distance between the sentence-sentence vectors produced for each sentence of the pair by the first neural network.

Learning thematic similarity metric from article text units

A method comprising: operating at least one hardware processor for: receiving, as input, at least one named entity, modifying said named entity based on a plurality of modification rules to generate a set of candidate named entities corresponding to said named entity, and identifying, for at least one candidate named entity in said set of candidate named entities, an article in a knowledge base of articles, wherein a title of said article matches said candidate named entity.

Automated entity disambiguation

A computer-implemented method, computerized apparatus and computer program product for claim generation, the method comprising: selecting at least one subject according to a given topic; selecting at least one verb from a first data source; selecting at least one object from a second data source; generating one or more candidate claim sentences, each of which composed of a subject selected from the at least one subject, a verb selected from the at least one verb and an object selected from the at least on object; and determining validity of the candidate claim sentences using a machine learning process.

Claim generation

A method comprising using at least one hardware processor for: receiving a topic under consideration (TUC) and a set of claims referring to the TUC; identifying semantic similarity relations between claims of the set of claims; clustering the claims into a plurality of claim clusters based on the identified semantic similarity relations, wherein said claim clusters represent semantically different claims of the set of claims; and generating a list of non-redundant claims comprising said semantically different claims.

Semantic merge of arguments

A method comprising using at least one hardware processor for receiving a topic under consideration (TUC); providing the TUC as input to a claim function, wherein the claim function is configured to mine at least one content resource, and applying the claim function to the at least one content resource, to extract claims with respect to the TUC; and providing the TUC as input to a classification function, and applying the classification function to one or more claims of the extracted claims, to output corresponding one or more classification tags, wherein each classification tag is associated with its corresponding claim.

Automatic construction of arguments

Growing community of inventors

Jerusalem, Israel

Noam Slonim