International Business Machines Corporation

Eyal Shnarch

Noam Slonim

Ajay Gupta

Amitkumar Manoharrao Paradkar

Ranit Aharonov

Yishai Abraham Feldman

Harshit Kumar

Amar Prakash Azad

Ran Levy

Prateeti Mohapatra

Suranjana Samanta

Lena Dankin

Alon Halfon

Ariel Gera

Leshem Choshen

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

Methods, systems, and computer program products for mining multi-party collaboration platforms to create triaging trees and playbooks are provided herein. A computer-implemented method includes obtaining, from a multi-user platform, conversations related to at least one technical issue; generating a plurality of triaging trees by analyzing the conversations, wherein each of the triaging trees stores information corresponding to temporal sequences of steps related to diagnosing and resolving said at least one technical issue; and deriving a playbook for resolving said at least one technical issue at least in part by combining two or more of the plurality of triaging trees.

Mining multi-party collaboration platforms to create triaging trees and playbooks

A computer-implemented method includes identifying at least one parse tree. The method includes identifying a pattern library. The method includes searching the pattern library for patterns that match at least one fragment of any of the at least one parse tree. The method includes determining whether the at least one parse tree is fully matched by a combination of matching patterns from the pattern library. The method includes ranking the at least one parse tree based on an extent to which the at least one parse tree is fully matched by the combination of matching patterns from the pattern library.

Evaluating parse trees in linguistic analysis

There is provided a computer-implemented method for outputting one or more cross-layer patterns to identify a target semantic phenomenon in text, the method comprising: extracting, for each word of at least some words of each training text fragment of training text fragments designated as representing a target semantic phenomenon, feature-values defined by respective layers; statistically analyzing the feature-values identified for the training text fragments to identify one or more cross-layer patterns comprising layers representing a common pattern for the training text fragments, the common cross-layer pattern defining one or more feature-values of a respective layer of one or more words and at least another feature-value of another respective layer of another word; and outputting the identified cross-layer pattern(s) for identifying a text fragment representing the target semantic phenomenon.

Systems and methods for language feature generation over multi-layered word representation

Growing community of inventors

Tel Aviv, Israel

Eyal Shnarch