International Business Machines Corporation

Rizwan Dudekula

Gandhi Sivakumar

Sujoy Sett

Sarvesh S Patel

Kushal S Patel

Lalit Agarwalla

Purushothaman Kunnath Narayanan

Wesley McMillan Devine

Naeem Altaf

Silpi Dhua

Jayachandu Bandlamudi

Shreeya Joshi

Ross Davis

A method, apparatus and computer program for detecting natural language (NL) bias by a conversational system is described. Embodiments of the invention determine an NL bias in a set of training questions used to train a machine learning model used by the conversational system to select a user intent. Other embodiments of the invention determine an NL bias in a user question received by the conversation system as compared to the set of training questions. The NL bias causes the machine learning to preferentially associate user queries to a user intent. In respective embodiments, the system takes a corrective action to adjust the NL bias of the training questions or the user question.

Natural language bias detection in conversational system environments

A processor may receive a media frame collection. The media frame collection may include two or more frames. The processor may identify, from the media frame collection, a location of an entity. The processor may determine whether the location of the entity is within a threshold location in regard to the other two or more frames in the media frame collection. The processor may display the one or more frames that exceed the threshold location to a user.

Dynamic sampling for object recognition

An audio search system is configured to perform a native search of one or more audio input files in response to a search query. The audio search system is connected to a corpus of audio files representing words, syllables, and characters that may be found in an audio input file. The audio search system has a memory storing instructions and a processing device configured to execute the instructions to receive a search query for searching one or more audio input files, convert the search query into an audio search expression, identify one or more meta-tags in the audio search expression, select a machine learning model based on the one or more meta-tags, and use the machine learning model to search the one or more audio input files for segments of the audio input file that are results of the search query.

System and method for searching audio data

A method, system and computer program product are provided for implementing 5G network-based information exchange across Internet of Things (IoT) enabled vehicular devices for amplified dynamic security. A 5G enabled amplified security manager is provided to implement enhanced amplified dynamic security for parked vehicles. The 5G enabled amplified security manager is provided with each enabled parked vehicle providing a 5G ad-hoc network. Information is dynamically collected and exchanged between the 5G enabled amplified security managers. When abnormal behavior is detected based on analysis of the collected data, the abnormal behavior information is sent on the 5G ad-hoc network triggering protective actions in co-ordination with all the devices in the network and provides security amplification.

Implementing information exchange across IoT enabled vehicular devices for amplified dynamic security

A method, system and computer program product are provided for implementing enhanced cognitively guiding visually impaired users. A 5G enabled guide for visually impaired (5G EGFVI) apparatus is provided to implement enhanced cognitively guiding visually impaired users. The 5G EGFVI apparatus provides an appropriate map for the visually impaired based on Internet of Things (IoT) enabled signals. An optimal path is provided by identifying IoT enabled traffic signals.

Implementing cognitively guiding visually impair users using 5th generation (5G) network

A method, computer system, and a computer program product for identifying sections in a document based on a plurality of visual features is provided. The present invention may include receiving a plurality of documents. The present invention may also include extracting a plurality of content blocks. The present invention may further include determining the plurality of visual features. The present invention may then include grouping the extracted plurality of content blocks into a plurality of categories. The present invention may also include generating a plurality of closeness scores for the plurality of categories by utilizing a Visual Similarity Measure. The present invention may further include generating a plurality of Association Matrices on the plurality of categories for each of the received plurality of documents based on the Visual Similarity Measure. The present invention may further include merging the plurality of categories into a plurality of clusters.

Growing community of inventors

Bengaluru, India

Rizwan Dudekula