Frequently Asked Questions: Addendum for the event on 28 September 2018
Overview of the next “hackathon” on September 28, 2018
The next instalment of the InnoTech Law Hackathon will be held on September 28, 2018 (see the Schedule). This time, participants will be asked to compete in a more focused, fast paced “hack” event to address problems arising from the theme of accessing justice in the Belt and Road region. The competition format will take the form of a “mystery shopper exercise” (see responses to the last two questions below) – an innovative approach that should be a first for a law hackathon in Asia.
Essentially, the problem scenario, background information, and preparation materials will be released to participants around 1 week before the competition “Demo Day”, and the participants will develop their tech solutions to be presented to a judging panel on September 28, 2018. Part of the assessment would be based on how the solutions handle a series of pre-set queries from the judges. Each team would only find out what these “secret” queries are on the day of the competition, which may or may not be in a conversational format.
We aim to provide the participants with workshop training in relevant software development platform and off-site mentoring (i.e. office hour drop-ins). Subject to the information on this page, other information is provided in the general FAQ of this website.
Why you should join the “hack”?
We welcome participants with an interest in the law and technology field, and legal and/or tech backgrounds to join. We will admit teams of 2-6 with a mix of such backgrounds into the competition. Among the team members there should be at least one or two members with beginner level coding or website development skills. One of our key aims is to make working with legaltech accessible and inclusive, and the development of relevant solutions highly feasible in a short period of time.
What are the substantive objectives?
The importance of (human) language and linguistics in the law cannot be overstated. Natural language processing (NLP) technologies play a critical role in bridging human and machine languages. The NLP framework will therefore feature prominently in the hack solutions in this event. The participants will learn and further develop their skills in operating in this area of computational law, which will gain increasing importance in future tech-enabled legal services.
Completion of this event will allow participants to create essential building blocks to either help launch new legaltech projects or improve their existing legaltech projects; whatever user interface (UI) or output form these projects should take. The participants should demonstrate creativity and subject matter expertise to help machines capture, extract, organize, and communicate legal meanings / human constructs – via verbal, visual, and other types of representation – and vice versa humans interacting with machines.
What are participants required to make as the tech “deliverables”?
Build your own “law bot”. The participants will make an “A.I. brain” and train it to help provide basic legal information: create simple legal concepts in the NLP framework (in the form of “intents” and “utterances”). The participants will develop these concepts to help the lay public and their “bots” search, extract and articulate legal meanings, and retrieve legal knowledge efficiently on a platform powered by the NLP engine.
Once the participants have created their prototype “A.I. brains”, they are free to deliver output by plugging in the digital “body parts” or modules of their choice, including chatbots, dashboards, optical character recognition (OCR), and other modules (some of these modules are provided by the technology sponsor Microsoft on the “LUIS” platform). The participants are encouraged to optimize the UI or user experience (UX) of their “digital body” interfaces for effective access by users.
The event is made for beginners: No software development or coding experience required, other than light use of APIs. Create your legal machine brain with Microsoft “ LUIS”, an online drag-and-drop platform with plenty of tools out-of-the-box to complete the requisite tasks in this event. We may also provide code samples and templates for the participants to integrate into their prototypes. Teams are also welcome to choose their own NLP framework.
What is your “law bot” meant to do, specifically?
Firstly, the participants are asked to build basic “A.I. brains” to address a simple legal scenario and issues arising from it. For example, the participants might be asked to respond to queries relating to probate law. A team might decide to address these problems by developing a chatbot. To build the “brain” of the chatbot, the participants will create a list of “intents” and “utterances” in a software framework powered by NLP. The NLP-based “A.I. brain” is essentially a knowledge graph that maps out concepts, I/O triggers, and information repositories to form a highly accessible and interactive “search engine” (in a broad sense).
Secondly, the participants will need some time to test their adapted framework to “reason” and process queries on simple legal concepts via various input and output (I/O) models: e.g. verbal Q&A, OCR queries etc. The ways to address the problems with the NLP engine could involve a variety of tech-enabled processes, including sentiment and search analytics and dashboard classifications. In “LUIS”, access to the bundle of “intents” and “utterances” would be published as application ID keys and plug-ins.
Thirdly, the participants will iteratively refine the “intents” and “utterances” to improve the quality of the machine responses. Each team should consider how to frame the user-machine interaction and reduce queries and corresponding responses into machine processes appropriate to the context of the problem scenario to optimize effectiveness and efficiency of their solutions.
Fourthly, the participants will present the solutions on “Demo Day” and be prepared to have the judges interact with each team’s “law bot” to test its abilities to render appropriate responses that would be helpful to a lay user accessing appropriate legal resources. Under the above example, on “Demo Day”, a judge could ask the chatbot, “how would I make a probate application?” The chatbot could then provide directions to the probate office at the High Court and other relevant information such as freely accessible online information.
Fifthly, how well each team does will depend on the relevancy and quality of the responses as well as the UI / UX. Please see the Judging Criteria for details.