Knowledgebra: An Algebraic Learning Framework for Knowledge Graph
نویسندگان
چکیده
Knowledge graph (KG) representation learning aims to encode entities and relations into dense continuous vector spaces such that knowledge contained in a dataset could be consistently represented. Dense embeddings trained from KG datasets benefit variety of downstream tasks as completion link prediction. However, existing embedding methods fell short provide systematic solution for the global consistency representation. We developed mathematical language based on an observation their inherent algebraic structure, which we termed Knowledgebra. By analyzing five distinct properties, proved semigroup is most reasonable structure relation general graph. implemented instantiation model, SemE, using simple matrix semigroups, exhibits state-of-the-art performance standard datasets. Moreover, proposed regularization-based method integrate chain-like logic rules derived human training, further demonstrates power language. As far know, by applying abstract algebra statistical learning, this work develops first formal graphs, also sheds light problem neural-symbolic integration perspective.
منابع مشابه
an investigation about the appropriate stochastic modeling framework for agricultural insurance pricing
با توجه به اینکه بیمه محصولات کشاورزی در ایران بیشتر جنبه ای حمایتی دارد و خسارات گزارش شده عموما بیش از حق بیمه های دریافت شده است، در این پایان نامه به جهت تعیین قیمت بیمه محصولات کشاورزی (گندم دیم) از فرآیندهای نوفه شلیک به عنوان مدلی مناسب استفاده شده است. بر اساس داده های صندوق بیمه کشاورزی از خسارات اعلام شده در سال زراعی 1388-1389 گندم دیم، در این پایان نامه حق بیمه خالص و ناخالص این محص...
Personal Knowledge/Learning Graph
Educational data mining and learning analytics have to date largely focused on specific research questions that provide insight into granular interactions. These insights have been abstracted to include the development of predictive models, intelligent tutors, and adaptive learning. While there are several domains where holistic or systems models have provided additional explanatory power, work...
متن کاملAN ALGEBRAIC STRUCTURE FOR INTUITIONISTIC FUZZY LOGIC
In this paper we extend the notion of degrees of membership and non-membership of intuitionistic fuzzy sets to lattices and introduce a residuated lattice with appropriate operations to serve as semantics of intuitionistic fuzzy logic. It would be a step forward to find an algebraic counterpart for intuitionistic fuzzy logic. We give the main properties of the operations defined and prove som...
متن کاملAn Ontology Based Knowledge Experiential Learning Framework
Ontology’s are no longer just a means of relating or unfolding semantic operations instead it is firmly contained within the virtual world and technologies have unravelled its full semantic potentials. The contribution of knowledge management combined with different software applications calls for increasing pedagogical support for extending both ontology and the knowledge management environmen...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Machine learning and knowledge extraction
سال: 2022
ISSN: ['2504-4990']
DOI: https://doi.org/10.3390/make4020019