Sri International

Yi Yao

Ajay Divakaran

Giedrius T Burachas

Arijit Ray

Karan Sikka

Xiao Lin

Michael A Cogswell

Yunye Gong

Rakesh Kumar

Supun Samarasekera

Girish Acharya

Han-Pang Chiu

John James Byrnes

Dayne B Freitag

Zachary Seymour

Niluthpol C Mithun

Michael Wessel

Jesse Hostetler

Anirban Roy

Louise Yarnall

Pallabi Ghosh

Kamran Alipour

Pritish Sahu

Meng Ye

Nikoletta Basiou

Stephanie Nunn

Sara Rutherford-quach

Zachary Weiler

A method, apparatus and system for adapting a pre-trained network for application to a different dataset includes arranging at least two different types of active adaptation modules in a pipeline configuration, wherein an output of a previous active adaptation module produces an input for a next active adaptation module in the pipeline in the form of adapted network data until a last active adaptation module, and wherein each of the at least two different types of adaptation modules can be switched on or off, determining at least one respective hyperparameter for each of the at least two different types of active adaptation modules, and applying the at least one respective determined hyperparameter to each of the at least two different types of active adaptation modules for processing received data from the pretrained network to determine an adapted network.

Modular adaptation for cross-domain few-shot learning

A method, apparatus and system for object detection in sensor data having at least two modalities using a common embedding space includes creating first modality vector representations of features of sensor data having a first modality and second modality vector representations of features of sensor data having a second modality, projecting the first and second modality vector representations into the common embedding space such that related embedded modality vectors are closer together in the common embedding space than unrelated modality vectors, combining the projected first and second modality vector representations, and determining a similarity between the combined modality vector representations and respective embedded vector representations of features of objects in the common embedding space to identify at least one object depicted by the captured sensor data. In some instances, data manipulation of the method, apparatus and system can be guided by physics properties of a sensor and/or sensor data.

Physics-guided deep multimodal embeddings for task-specific data exploitation

A computing system comprising a memory configured to store an artificial intelligence (AI) model and an image, and a computation engine executing one or more processors may be configured to perform the techniques for error-based explanations for AI behavior. The computation engine may execute the AI model to analyze the image to output a result. The AI model may, when analyzing the image to output the result, process, based on data indicative of the result, the image to assign an error score to each image feature extracted from the image, and obtain, based on the error scores, an error map. The AI model may next update, based on the error map and to obtain a first updated image, the image to visually indicate the error score assigned to each of the image features, and output one or more of the error scores, the error map, and the first updated image.

Error-based explanations for artificial intelligence behavior

A method, apparatus and system for determining user-content associations for determining and providing user-preferred content using multimodal embeddings include creating an embedding space for multimodal content by creating a first modality vector representation of the multimodal content having a first modality, creating a second modality vector representation of the multimodal content having a second modality, creating a user vector representation, as a third modality, for each user associated with at least a portion of the multimodal content, and embedding the first and the second modality vector representations and the user vector representations in the common embedding space using at least a mixture of loss functions for each modality pair of the first, the at least second and the third modalities that pushes closer co-occurring pairs of multimodal content. Embodiments can further include generating content using determined attributes of a message to be conveyed and features of the user-preferred content.

User targeted content generation using multimodal embeddings

This disclosure describes machine learning techniques for capturing human knowledge for performing a task. In one example, a video device obtains video data of a first user performing the task and one or more sensors generate sensor data during performance of the task. An audio device obtains audio data describing performance of the task. A computation engine applies a machine learning system to correlate the video data to the audio data and sensor data to identify portions of the video, sensor, and audio data that depict a same step of a plurality of steps for performing the task. The machine learning system further processes the correlated data to update a domain model defining performance of the task. A training unit applies the domain model to generate training information for performing the task. An output device outputs the training information for use in training a second user to perform the task.

Machine learning system for technical knowledge capture

A method, apparatus and system for training an embedding space for content comprehension and response includes, for each layer of a hierarchical taxonomy having at least two layers including respective words resulting in layers of varying complexity, determining a set of words associated with a layer of the hierarchical taxonomy, determining a question answer pair based on a question generated using at least one word of the set of words and at least one content domain, determining a vector representation for the generated question and for content related to the at least one content domain of the question answer pair, and embedding the question vector representation and the content vector representations into a common embedding space where vector representations that are related, are closer in the embedding space than unrelated embedded vector representations. Requests for content can then be fulfilled using the trained, common embedding space.

System and method for content comprehension and response

Techniques are disclosed for identifying multimodal subevents within an event having spatially-related and temporally-related features. In one example, a system receives a Spatio-Temporal Graph (STG) comprising (1) a plurality of nodes, each node having a feature descriptor that describes a feature present in the event, (2) a plurality of spatial edges, each spatial edge describing a spatial relationship between two of the plurality of nodes, and (3) a plurality of temporal edges, each temporal edge describing a temporal relationship between two of the plurality of nodes. Furthermore, the STG comprises at least one of: (1) variable-length descriptors for the feature descriptors or (2) temporal edges that span multiple time steps for the event. A machine learning system processes the STG to identify the multimodal subevents for the event. In some examples, the machine learning system comprises stacked Spatio-Temporal Graph Convolutional Networks (STGCNs), each comprising a plurality of STGCN layers.

Identifying complex events from hierarchical representation of data set features

In general, the disclosure describes various aspects of techniques for attention-based explanations for artificial intelligence behavior. A device comprising a memory and a computation engine executing a processor may be configured to perform the techniques. The memory may store the artificial intelligence model and the image. The computation engine may receive a query regarding the image, and execute the artificial intelligence model to analyze the image in order to output the result to the query. The artificial intelligence model may, when analyzing the image to output the result, segment the image into hierarchically arranged semantic areas in which objects in the image are segmented into parts, determine, based on the query, an attention mask for the areas, update, based on the attention mask, the image to visually identify which of the areas formed a basis for the result, and output the updated image.

Attention-based explanations for artificial intelligence behavior

Growing community of inventors

Princeton, NJ, United States of America

Yi Yao