Wrapping up 2016 and wishing all of you a happy 2017!
As the year 2016 comes to a close, we on behalf of NLPCORE would like to wish you and your loved ones a very happy holiday season and 2017. We also take this opportunity to thank you for your support, feedback and contributions in helping us refine our thinking and improve upon our solutions stack. Here is a brief update on NLPCORE through its trials, tribulations and eventual triumphs during 2016 and plans on anvil for a bright future ahead in 2017!
We began 2016 with a number of hardware experiments with a goal to find an optimized hardware stack across CPU, GPU, RAM, Storage and network latency that helps us minimize cost without compromising on performance. If you haven't already, please read about our experiments here.
While getting the infrastructure on the right course, Varun also focused on core algorithms and working with Alexander Ratushny (Fred Hutch and now Celgene) we validated prowess of our search engine against manually verified gene-virus interactions. Not only we fared well on Precision/Recall numbers, we also were able to identify interactions buried deep in research papers that were not manually verified in an otherwise comprehensive study. Read more about our application note draft here. We will be revising it next year with our latest algorithms and hardware infrastructure.
With improved infrastructure and core algorithms, and great feedback from you, we put together our engineering requirements specification (see it here). With a few missteps and failed internships behind us, we are very fortunate to have partnered with dmmd.net who took our requirements document, proof of concept implementation at nlpcore.com and our new APIs as raw materials and have been busy converting it into a polished product. This hands-off third party development approach also has helped us refine our platform so that it is not only well-defined for re-use and customizations but also that it is responsive and scalable! You may read more on our engineering process innovations here.
With 2016 coming to close, we have outlined the followings as our priorities ahead:
Firstly, we are getting our Lifesciences Search, Collaboration and Procurement solution pilot ready. Our site will provide users ability to explore bioentities, reagents, see their linkages, interaction references into various articles - both through document and graph views. We also intend to provide users ability to define their own categories or entity types (such as methods, test subjects, and gender).
We will be documenting all our platform APIs, writing a number of samples up and down the stack - be it: how to add a new document/web link for search or add a new technique to identify an entity (eg DNA sequences) or add and expose optimization parameters for a search technique or add another view in our user-interface. We hope to attract developers with wide range of skills - math, java, python, front-end, infrastructure, algorithms...
We will also be doing performance analysis for our search results and entity extractions and revising our draft application note based on new findings. Our new algorithms start with the exhaustive data set and therefore we hope to improve both on precision and recall fronts.
Last but not the least, our focus in the latter half of the year will be on securing customer deployments, improving page footprints and generating revenues. In doing so, we aim for 2018 to be our transition from an early pre-revenue startup to a viable business enterprise.