Abstract: A method, system, and non-transitory computer-readable medium can include generating a prompt comprising input data. The prompt can be configured to process textual data, numerical data, mixed data, or combinations thereof. The input data can include at least one of new data, a selection of historical samples from historical data, or anomalies in event data from a stream. The prompt can then be provided to a large language model (LLM), which generates a synthetic label based on the prompt. The method can then determine if a prediction model is drifting by comparing a prediction label against the synthetic label for a match. The prediction model can generate the prediction label based on the input data. If the prediction label and the synthetic label do not match, the method can include initiating a corrective action. The corrective action can include at least one of generating an alert for a user, retraining the prediction model, or initiating an investigation.

Abstract: Methods, network traffic manager apparatuses, non-transitory computer readable media, and systems that optimize a representation for detecting a pattern of a traffic flow are disclosed. The method includes retrieving a representation and dataset associated with the representation from storage and prompting a natural language processing model to transform the retrieved representation to a first candidate representation different than the retrieved representation based on the dataset. Then the method includes inputting the first candidate representation into a simulator to generate one or more first candidate metric values and evaluating whether to replace the retrieved representation with the first candidate representation based on the generated one or more first candidate metric values. This process may be iteratively implemented to optimize the generated candidate representation.

Abstract: Methods, network traffic management apparatuses, non-transitory computer readable media, and systems that optimize transport of unknown and changing unstructured log data in a network environment.