HAKE: A Knowledge Engine Foundation for Human Activity Understanding

نویسندگان

چکیده

Human activity understanding is of widespread interest in artificial intelligence and spans diverse applications like health care behavior analysis. Although there have been advances with deep learning, it remains challenging. The object recognition-like solutions usually try to map pixels semantics directly, but patterns are much different from patterns, thus hindering another success. In this work, we propose a novel paradigm reformulate task two-stage: first mapping an intermediate space spanned by atomic primitives, then programming detected primitives interpretable logic rules infer semantics. To afford representative primitive space, build knowledge base including 26+ M labels human priors or automatic discovering. Our framework, H xmlns:xlink="http://www.w3.org/1999/xlink">uman xmlns:xlink="http://www.w3.org/1999/xlink">A xmlns:xlink="http://www.w3.org/1999/xlink">ctivity xmlns:xlink="http://www.w3.org/1999/xlink">K xmlns:xlink="http://www.w3.org/1999/xlink">nowledge xmlns:xlink="http://www.w3.org/1999/xlink">E xmlns:xlink="http://www.w3.org/1999/xlink">ngine ( xmlns:xlink="http://www.w3.org/1999/xlink">HAKE ), exhibits superior generalization ability performance upon canonical methods on challenging benchmarks. Code data available at http://hake-mvig.cn/ .

برای دانلود باید عضویت طلایی داشته باشید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Experimental Semiotics: an Engine of Discovery for Understanding Human Communication

The recent growth of Experimental Semiotics (ES) offers us a new option to investigate human communication. We briefly introduce ES, presenting results from three themes of research which emerged within it. Then we illustrate the contribution ES can make to the investigation of human communication systems, particularly in comparison with the other existing options. This comparison highlights ho...

متن کامل

A New Foundation For Knowledge Systems

We propose mathematical foundations for knowledge systems based on the notion of effectively presented domain. The paper aims to give a schematic model for computer’s behavior in changing information environment. In this paper, we focus on computational resources rather than on computer’s particular actions.

متن کامل

Towards a cognitive foundation for knowledge representation

Knowledge engineering, knowledge management and conceptual modelling are concerned with representing knowledge of business and organizational domains. These research areas use ontologies for knowledge representation. Ontologies are understood either in the philosophical sense as firm metaphysical commitments or in the looser sense of dictionaries or taxonomies. This paper critically examines th...

متن کامل

A Foundation for Hybrid Knowledge Bases

Hybrid knowledge bases (HKB’s) [11] were developed to provide formal models for the mediation of data and knowledge bases [14,15]. They are based on Generalized Annotated Logic Programming (GAP)[7] and employ an inference mechanism, HKB-resolution, that is considerably simpler than those that have been proposed for GAP. The simplicity of HKB-resolution is explained in this paper by showing that...

متن کامل

ذخیره 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: IEEE Transactions on Pattern Analysis and Machine Intelligence

سال: 2023

ISSN: ['1939-3539', '2160-9292', '0162-8828']

DOI: https://doi.org/10.1109/tpami.2022.3232797