Amazon Technologies, Inc.

David Anthony Leen

Christopher Wiswall Greene

Kevin Michael McCormick

Sravan Babu Bodapati

Alex Kharbush

Techniques for phased deployment of machine learning models are described. Customers can call a training API to initiate model training, but then must wait while the training completes before the model can be used to perform inference. Depending on the type of model, machine learning algorithm being used for training, size of the training dataset, etc. this training process may take hours or days to complete. This leads to significant downtime where inference requests cannot be served. Embodiments improve upon existing systems by providing phased deployment of custom models. For example, a simple, less accurate model, can be provided synchronously in response to a request for a custom model. At the same time, one or more machine learning models can be trained asynchronously in the background. When the machine learning model is ready for use, the customers' traffic and jobs can be transferred over to the better model.

Phased deployment of deep-learning models to customer facing APIs

A machine learning-based service processes requests using compressed and complete models to provide faster response times for servicing requests to process data. Initially, a host processes data using a compressed model that is stored in the host's memory and then switches to a larger, more accurate complete model after it is loaded into the host's memory. A host of the machine-learning based service may receive one or more requests to process data. In response, the host uses a compressed version of a model to begin processing the data. The host starts loading the complete version of the model into the host's memory. When the complete version of the model is loaded into memory, the host switches to process a remaining portion of the data using the complete version of the model.

Processing requests using compressed and complete machine learning models

Techniques for query-based data modification are described herein. In some examples, a first search of searchable data may be performed based, at least in part, on a first query. A second search of the searchable data may be performed based, at least in part, on a second query. One or more results of the first search may be provided. Input may be received for associating a first result of the first search with the second search. At least a first modification, related to the input, to perform on the searchable data may then be determined. The first modification may then be performed on the searchable data.

Query-based data modification

Techniques for parsing and execution of data including multiple information classes are described herein. In some examples, a collection of data may include multiple information classes through which the data may be parsed and analyzed. In some examples, the multiple information classes may include a textual character information class, a visual style information class, and an inferred information class, such as may include data identifiable based on information external to the data collection. A plurality of tokens associated with the data collection may be generated. One or more of the plurality of tokens may be organized into a set of instructions. The set of instructions may be provided to a computer program for execution.

Growing community of inventors

Seattle, WA, United States of America

David Anthony Leen