Wells Fargo Bank, N.a.

Manesh Saini

Hasan Adem Yilmaz

Naveen Gururaja Yeri

Srimoyee Duttagupta

Pranshu Sharma

Hemalatha Ac

Alan On Yau

Dipan Banerjee

Michelle Sunna Nowe

Carleton J Lindgren

Jeff J Stapleton

Michael Knorr

Sadie Salim

Austin Lee Grelle

Michael Paul Bortis

Narsingh Rao Chatla

Ronald Louis Sobey

Pradyut K Parida

Yashjeet Singh

Fanus Arefaine

Kamila Rywelska

Anara Myrzabekova

Siddharth Jain

Guangcao Ji

Winthrop Treynor Smith

Disclosed herein is a workflow for a chatbot system based on an ad hoc set of documents. The chatbot enables users to ask questions of these documents. The workflow then searches for relevant information and generates a response. The response may include an answer to a question and a relevant section of a document.

Document set interrogation tool

Systems, apparatuses, methods, and computer program products are disclosed for constructing a foundational identity to establish the identity of an individual and linking the foundational identity to various other identities of the individual in virtual spaces. An example method includes linking identities across virtual spaces. The example method further includes verifying identity activity associated with the various other identities using the link between the various other identities and the foundational identity, and based on the verification of the identity activity, generate outputs.

Systems and methods for identifying activities of individuals in virtual spaces using identity linking

Quality associated with an interpretation of data captured as unstructured data can be determined. Attributes can be identified within the unstructured data automatically. Subsequently, sentiment associated with each of the attributes can be determined based on the unstructured data. Correctness of the unstructured data, and thus the interpretation, can be assessed based on a comparison of the attribute and associated sentiment with structured data. A quality score can be generated that captures the quality of the data interpretation in terms of correctness and as well as results of another analysis including completeness, among others. Comparison of the quality score to a threshold can dictate whether or not the interpretation is subject to further review.

Data interpretation analysis

Disclosed is an approach for evaluating internal consistency of object classifications using machine learning modeling. In an example, models are iteratively trained, using supervised learning, on different majority segments (e.g., about 90%) of a dataset as training data segments. Trained models can be applied to the remaining data (e.g., about 10%) as test data segments to obtain, for each object, a predicted classification and a confidence score. Models in training iterations (e.g., 10 iterations) may be independently trained on substantially non-overlapping test data segments (with each iteration testing, e.g., about 10% of the dataset). When a model's predicted classification disagrees from a prior classification, and the confidence of the prediction is sufficiently high (indicating sufficiently strong disagreement), the object's prior classification may be revised. Training data, other than the data itself being evaluated for consistency, is not necessarily required.

Iteratively trained machine learning models for evaluations of internal consistency

Risk assessment can be performed in many contexts including in lending. A set of metrics can be received and derived from data associated with a party, such as a third-party retailer or dealer, with respect to performance of an activity. A subset of metrics can be identified that exceed a threshold of acceptable performance. Weights can be applied to at least the subset of metrics that captures significance of corresponding metrics. A single weighted score can be computed from aggregation of the weighted subset of metrics, and a third party can be classified based on comparison of the weighted score to a predetermined threshold.

Risk assessment in lending

Disclosed are systems and method of mapping data entries originating in different systems. A plurality of data entries from different systems are normalized such that they can be compared to each other and mapped, even though the data entries are defined by data fields with differing phrases, descriptive details, and lengths of detail. Data entries may be filtered according to data fields before a mapping operation is employed for mapping. The mapping operation evaluates similarity scores based on the data fields using a combination of exact matching algorithms, dictionary matching algorithms, and text mining algorithms. The mapped data entries and data fields are displayed to a user.

Mapping disparate datasets

Growing community of inventors

New York, NY, United States of America

Manesh Saini