Knowledge bases (KB) facilitate real world decision making by providing access to structured relational information that enables pattern discovery and semantic queries. Although there is a large amount of data available for populating a KB; the data must first be gathered and assembled. Traditionally, this integration is performed automatically by storing the output of an information extraction pipeline directly into a database as if this prediction were the ``truth.'' However, the resulting KB is often not reliable because (a) errors accumulate in the integration pipeline, and (b) they persist in the KB even after new information arrives that could rectify these errors. We envision a paradigm-shift in KB construction for addressing these concerns that we term an ``epistemological'' database. In epistemological databases the existence and properties of entities are not directly input into the DB; they are instead determined by inference on raw evidence input into the DB. This shift in thinking is important because it allows inference to revisit previous conclusions and retroactively correct errors as new evidence arrives. Evidence is abundant and in steady supply from web spiders, semantic web ontologies, external databases, and even groups of enthusiastic human editors. As this evidence continues to accumulate and inference continues to run in the background, the quality of the knowledge base continues to improve. In this dissertation we develop the machine learning components necessary to achieve epistemological knowledge base construction at scale with key contributions in modeling, inference and learning.

Sarah Moss argues that in addition to full beliefs, credences can constitute knowledge. She introduces the notion of probabilistic content and shows how it plays a central role not only in epistemology, but in the philosophy of mind and language. Just you can believe and assert propositions, you can believe and assert probabilistic contents.

Problems can arise whenever inferencing is attempted on a knowledge base that is incomplete. Our work shows that data mining techniques can be applied to fill in incomplete areas in Bayesian Knowledge Bases (BKBs), as well as in other knowledge-based systems utilizing probabilistic representations. The problem of inconsistency in BKBs has been addressed in previous work, where reinforcement learning techniques from neural networks were applied. However, the issue of automatically solving incompleteness in BKBs has yet to be addressed. Presently, incompleteness in BKBs is repaired through the application of traditional knowledge acquisition techniques. We show how association rules can be extracted from databases in order to replace excluded information and express missing relationships. A methodology for incorporating those results while maintaining a consistent knowledge base is also included.

Excerpt from Experience and Prediction: An Analysis of the Foundations and the Structure of Knowledge The conviction that the key to an understanding of sci entific method is contained within the probability problem grew stronger and stronger with me in the face of such basic mistakes. This is the reason why, for a long time, I renounced a comprehensive report of my epistemological views, although my special investigations into different problems of epistemology demanded a construction of foundations different from those constructed by some of my philosophical friends. I concentrated my inquiry on the problem of probability which demanded at the same time a mathematical and a logical analysis. It is only after having traced out a logistic theory of probability, includ ing a solution of the problem of induction, that I turn now to an application of these ideas to questions of a more gen eral epistemological character. As my theory of probabil ity has been published for some years, it was not necessary to present it with all mathematical details once more in the present book; the fifth chapter, however, gives an ah breviated report of this theory - a report which seemed necessary as the probability book has been published in German only. About the Publisher Forgotten Books publishes hundreds of thousands of rare and classic books. Find more at www.forgottenbooks.com This book is a reproduction of an important historical work. Forgotten Books uses state-of-the-art technology to digitally reconstruct the work, preserving the original format whilst repairing imperfections present in the aged copy. In rare cases, an imperfection in the original, such as a blemish or missing page, may be replicated in our edition. We do, however, repair the vast majority of imperfections successfully; any imperfections that remain are intentionally left to preserve the state of such historical works.

This work breaks new ground by carefully distinguishing the concepts of belief, confirmation, and evidence and then integrating them into a better understanding of personal and scientific epistemologies. It outlines a probabilistic framework in which subjective features of personal knowledge and objective features of public knowledge have their true place. It also discusses the bearings of some statistical theorems on both formal and traditional epistemologies while showing how some of the existing paradoxes in both can be resolved with the help of this framework.This book has two central aims: First, to make precise a distinction between the concepts of confirmation and evidence and to argue that failure to recognize this distinction is the source of certain otherwise intractable epistemological problems. The second goal is to demonstrate to philosophers the fundamental importance of statistical and probabilistic methods, at stake in the uncertain conditions in which for the most part we lead our lives, not simply to inferential practice in science, where they are now standard, but to epistemic inference in other contexts as well. Although the argument is rigorous, it is also accessible. No technical knowledge beyond the rudiments of probability theory, arithmetic, and algebra is presupposed, otherwise unfamiliar terms are always defined and a number of concrete examples are given. At the same time, fresh analyses are offered with a discussion of statistical and epistemic reasoning by philosophers. This book will also be of interest to scientists and statisticians looking for a larger view of their own inferential techniques.The book concludes with a technical appendix which introduces an evidential approach to multi-model inference as an alternative to Bayesian model averaging.