The workshop entitled "Fragile Earth: Theory Guided Data Science to Enhance Scientific Discovery" (in short, the "FEED Workshop") is a full-day workshop to be held in conjunction with the prestigious ACM KDD 2018 (Knowledge Discovery and Data mining) conference. The KDD conference covers data science (including machine learning [ML] aspects of Artificial Intelligence [AI]), is highly selective, and has two tracks: Research Track and Applied Data Science Track. In addition, KDD organizers invite workshop proposals to cater to new and emerging areas. Accepted workshop proposals for this year (including this full-day workshop) are listed on the KDD 2018 workshop page. The 2018 KDD FEED workshop webpage is hosted online by the AI for Good Foundation. Northeastern's Sustainability and Data Sciences Laboratory (SDS Lab) PhD student Thomas J. ("TJ") Vandal and his adviser CEE Professor Auroop R. Ganguly, in colloaboration with NASA Ames, SDS Lab spinout risQ Inc. (Northeastern SDS Lab alum Dr. Evan Kodra) and others, have been presenting at both the regular KDD conferences and (Ganguly has been) organizing workshops. The KDD 2017 runner-up best paper award in the Applied Data Science Track was won by their "DeepSD paper" on Deep Learning for Statistical Downscaling in climate. The KDD 2018 conference, in their Research Track, accepted their paper on Uncertainty Quantification with Bayesian Deep Learning, applied to DeepSD (this paper is currently available on the arXiv preprint server here). The 2018 KDD paper includes Northeastern ECE Professor Jennifer Dy as a co-author. Ganguly has a long association with KDD conferences, and he was the founding organizer of six successive workshops on Knowledge Discovery from Sensor Data (SensorKDD), starting with the first such workshop in 2007 which led to an edited book on the topic.