International Business Machines Corporation

Ateret Anaby-Tavor

David Amid

David Boaz

Ofer Shir

Inbal Ronen

Amir Kantor

Ofer Lavi

Salil Ahuja

Ella Rabinovich

Segev Shlomov

Mitchell Mason

Rama K T Akkiraju

Guy Uziel

Shahar Chen

Rotem Aharonov

Jason M Leonard

Jalal U Mahmud

Thomas Guthrie Zimmerman

Vibha Singhal Sinha

Rama Kalyani T Akkiraju

Oded Margalit

Peter Bak

Danish Contractor

Boaz Carmeli

Sergey Zeltyn

Lior Limonad

Jerald Thomas Schoudt

Maria Nichole Schwenger

Gaurav Pandey

Dmitry Moor

Ashraf Haib

Matan Vetzler

Koren Ran Lazar

Moran Gavish

A computer-implemented method, system and computer program product for identifying a semantic representation for a set of texts. The set of texts is encoded using a language model to obtain a set of corresponding texts' contextualized embeddings. A centroid of the contextualized embeddings is then calculated. A user-designated number of words from a pre-defined vocabulary of the language model that have their non-contextualized embeddings closest to the centroid are then identified. Furthermore, permutations of the identified words are calculated using an n-gram range. The permutations are then encoded using the language model in a contextualized manner. The encoded permutation from the encoded permutations with the greatest similarity to the centroid is identified. The identified encoded permutation is then assigned as the semantic representation of the set of texts. In this manner, the semantic representation is more effectively identified by expanding the embedding space upon which the semantic representation is chosen.

Identifying a semantic representation for texts by expanding the embedding space

An example system includes a processor to receive a bot design and escalation logs associated with a chat bot configured based on the bot design. The processor can compute a similarity score between each of a number of bot response nodes in the bot design and the escalation logs. The processor can generate a sorted list of the bot response nodes in the bot design based on the similarity scores.

Adjusting chat bot design based on matched escalation logs

A method, computer system, and a computer program product for automated agent intent detection enhancement are provided. A first message from a first user is received. The first message is generated during a first conversation between the first user and a first automated agent. A computer produces a second message that includes a same request as the first message but a different language modality than the first message. The second message and the first message are combined to form a combined message. The combined message is input into the first automated agent such that the first automated agent produces an intent classification for the first message.

Message mapping and combination for intent classification

Methods, systems, and computer program products for implementing automated communication exchange programs for attended robotic process automation are provided herein. A computer-implemented method includes invoking, during a user communication associated with an attended robotic process automation context, at least one automated communication exchange program in response to at least one user input; determining, using the at least one automated communication exchange program, information directed to the at least one user input; carrying out, using the at least one automated communication exchange program, at least a portion of the user communication subsequent to determining the information directed to the at least one user input; and performing one or more automated actions in connection with automatically carrying out the at least a portion of the user communication.

Automated communication exchange programs for attended robotic process automation

An example system includes a processor to receive data for a multi-objective task. The processor is to also perform the multi-objective task on the received data via a trained primal network. The primal network and a dual network are trained for a multi-objective task using a Lagrangian loss function representing a number of objectives. The primal network is trained to minimize the Lagrangian loss function and the dual network is trained to maximize the Lagrangian loss function.

Performing multi-objective tasks via primal networks trained with dual networks

An intermediary server receives a resource request from a client computer for a requested resource that has a desired resource ability to optimize the client computer. The intermediary server identifies an alternative resource that has the desired resource ability to optimize the client computer, and determines that an ability of the alternative resource to provide the desired resource ability exceeds an ability of the requested resource to provide the desired resource ability. In response to determining that the ability of the alternative resource to provide the desired resource ability exceeds the ability of the requested resource to provide the desired resource ability, the intermediary server instructs a resource server to send the alternative resource instead of the requested resource to the client computer.

Providing an optimal resource to a client computer via interactive dialog

An example system includes a processor to receive a prefix of conversation and a text input. The processor is to also generate a completed response based on the prefix of conversation and the text input via a trained primal network. The primal network is trained to minimize a Lagrangian loss function representing a number of objectives and a dual network is trained to maximize the Lagrangian loss function.

Generating completed responses via primal networks trained with dual networks

An intermediary server receives a resource request for a requested resource from a client computer. The intermediary server responds to the resource request by initiating an interactive dialog between the intermediary server and the client computer, where the interactive dialog identifies a desired resource ability of the requested resource. The intermediary server identifies, based on responses to the interactive dialog from the client computer, an alternative resource that best meets the desired resource ability, wherein the requested resource and the alternative resource are a same type of resource, and wherein the interactive dialog uses an interactive exchange that interprets interactive dialog responses from the client computer in order to identify the desired resource ability and the alternative resource that best meets the desired resource ability and then instructs a resource server to send the alternative resource to the client computer.

A method of selecting a group from a plurality of multi objective designs which comply with a plurality of objectives. The method comprises providing a plurality of multi objective designs, each the multi objective design having a plurality of multi objective design objective values which comply with at least one constraint of a Pareto Frontier of an objective space of a plurality of objectives, selecting a group from the plurality of multi objective designs, each member of the group is selected according to a match between at least one objective of respective the plurality of objectives and at least one of a respective gain threshold and a respective loss threshold, and outputting the group.

Multi objective design selection

A mechanism is provided in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to cause the at least one processor to implement a personalized interactive decision support system. A customer profile generator executing within the personalized interactive decision support system determines a set of traits for at least one customer based on at least one user-generated content source. A personalized product recommendation module executing within the personalized interactive decision support system matches the at least one customer to at least one product within a set of products based on a consumption preferences-to-product attribute mapping data structure to form at least one product recommendation. Each product in the set of products has an associated set of attributes and wherein the consumption preferences-to-product attribute mapping data structure correlates customer traits and consumption preferences and correlates consumption preferences and product attributes. A visual and interactive decision support module executing within the personalized interactive decision support system presents the at least one product recommendation.

Interactive decision support based on preferences derived from user-generated content sources

A method, apparatus and product for automated multi-objective solution selection. The method comprises obtaining a set of solutions to a multi-objective problem, wherein each solution is a non-dominated solution. The method further comprises computing outranking weights for each pair of solutions and selecting a subset of the set of solutions based upon a defensive criteria in with respect to the outranking weights, wherein the defensive criteria relates to a number of solutions that outrank the subset. The method further comprises outputting the subset of the set of solutions.

Automated multi-objective solution selection

Data storage management by determining, for leaf and summary storage spaces of a data storage space hierarchy having at least two leaf storage spaces descending from at least one summary storage space, an invariant leaf attribute value for each leaf attribute type, an invariant summary attribute value for each descending leaf attribute type as a sum of the invariant leaf attribute values of all leaves descending from the summary storage space, and for each leaf, a variable leaf attribute value for each leaf attribute type, and, for each summary storage space, a variable summary attribute value for each descending leaf attribute type, where for each summary storage space, and for each storage space immediately descending from the summary storage space, each variable leaf attribute value of the immediately descending storage space is expressed as a proportion of the variable summary attribute value for the same attribute type.

Efficient management of hierarchically-linked data storage spaces

A method comprising using at least one hardware processor for: receiving a multi-objective optimization problem; projecting a Pareto frontier of candidate solutions for said multi-objective optimization problem to a hyperplane; decomposing said hyperplane into multiple Voronoi regions each associated with a candidate solution of said candidate solutions; determining a robustness degree for each candidate solution of said candidate solutions, by computing a hypervolume for each region of said multiple Voronoi regions; and ranking said candidate solutions based on the robustness degree.

Objective weighing and ranking

A method of visualizing a plurality of designs which comply with a plurality of objectives. The method comprises acquiring a plurality of designs each represented by sequential multivariate data indicative of a compliance with a plurality of objectives, generating an objective anchored based self-organizing map (SOM) having a plurality of objective anchors and maps the plurality of designs in an objective space, and outputting the objective anchored based SOM. Each objective anchor is associated with one of the plurality of objectives, each the design is visualized in the objective anchored based SOM by an indicator which the distance thereof from each the objective anchors is indicative of a compliance thereof with a respective the associated objective in relation to other of the plurality of objectives.

Growing community of inventors

Givat Ada, Israel

Ateret Anaby-Tavor