ADOBE has filed a patent for techniques to identify bot activity. The patent describes a computer-implemented method that involves receiving click activity samples, classifying them using a machine learning model, filtering the data based on the classification predictions, and modifying the user interface accordingly. The method also includes training the machine learning model using labeled and unlabeled samples and a topological loss function. The trained model can be used to classify click activity data and identify bot activity. GlobalData’s report on ADOBE gives a 360-degree view of the company including its patenting strategy. Buy the report here.

According to GlobalData’s company profile on ADOBE, User behaviour analysis was a key innovation area identified from patents. ADOBE's grant share as of September 2023 was 75%. Grant share is based on the ratio of number of grants to total number of patents.

Techniques for identifying bot activity using machine learning

Source: United States Patent and Trademark Office (USPTO). Credit: ADOBE Inc

A recently filed patent (Publication Number: US20230316124A1) describes a computer-implemented method for classifying click activity data to identify bot activity. The method involves accessing a mixed plurality of samples that includes labeled samples of a first class (click activity by authenticated users) and unlabeled samples (click activity by unauthenticated users). For each sample, a classifier model calculates a corresponding class probability. A sample selection module then selects a training set of samples from the unlabeled samples based on the class probability. Using a topological loss function module, a machine learning model is trained to classify samples among the first and second classes. The trained model can be used to identify bot activity in click activity data.

The topological loss function used in the method is based on the distance between the topological signature of the input space and the topological signature of the latent space of the first class. It includes a regularization term based on this distance and a second distance between the topological signatures of the input and latent spaces of the second class.

The method also includes filtering the click activity data to exclude bot activity. This is done by generating filtering criteria based on the information from the classification predictions. The activity of bot users is then excluded from the filtered click activity data based on these criteria.

The patent also describes a system that implements the method. The system includes one or more processing devices and a non-transitory computer-readable medium. The processing devices execute program code stored in the medium to perform various operations, including receiving samples of click activity data, classifying the samples using a machine learning model, filtering the click activity data based on the classification predictions, and modifying a user interface based on the filtered click activity data.

In addition to identifying bot activity, the system can also perform clustering of the samples based on the classification predictions and calculate various statistics for each cluster. It can generate a graph that represents the clusters and their relationships. The system can further modify the user interface based on the values of these statistics.

Overall, this patent presents a computer-implemented method and system for classifying click activity data and identifying bot activity. The method utilizes a topological loss function and machine learning to achieve accurate classification. The system provides a means to filter and analyze click activity data, enabling the identification and exclusion of bot activity.

To know more about GlobalData’s detailed insights on ADOBE, buy the report here.

Premium Insights

From

The gold standard of business intelligence.

Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.

GlobalData

GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.

GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.