Abstract: The business problem of assigning task and activity codes to timecard narratives is challenging because the narrative text is relatively short and many different categories have similar narratives. Effective solutions require a combination of a sequence models and pattern matching. Pre-defined keywords and phrases can be used as additional features in data preparation or in a post-processing results.
Abstract: This session will detail how Fredrikson & Byron, a Minnesota-based firm of 300 lawyers has worked to reconcile two directly opposing goals: 1. Using task codes to leverage the firm’s historical data to generate accurate fee estimates 2. Allowing lawyers to have this data-driven approach without task code entry affecting their own workflows.
Abstract: Legal data is largely unstructured/untagged, and computation requires some structure — permitting communication across systems. SALI is a non-profit providing standardized fields/identifiers for 3,000+ legal items. SALI will (1) describe its standard, (2) explain how to use it, and (3) illustrate applications in both relational databases and graph. Connect systems and more-easily build upon prior research — all by harnessing the power of standards.
Abstract: The emergence of big data has given rise to Dataism – a mindset or philosophy that argues relying on data could reduce cognitive biases and illuminate patterns of behavior we haven't yet noticed. But what are the consequences of Dataism in the legal industry and how do we determine the applicability of AI solutions to augment service delivery?
Abstract: This session will show how our system improves the mortgage process followed by banks or real-estate agencies by including auto-classification of documents and Named Entity Recognition to convert unstructured data into structured information. Our product helps human operators to better understand the legal documents and it increases their capacity to process much bigger amounts of documents with the same resources.