Learning to Copy Coherent Knowledge for Response Generation

نویسندگان

چکیده

Knowledge-driven dialog has shown remarkable performance to alleviate the problem of generating uninformative responses in system. However, incorporating knowledge coherently and accurately into response generation is still far from being solved. Previous works dropped paradigm non-goal-oriented knowledge-driven dialog, they are prone ignore effect goal, which potential impacts on exploitation generation. To address this problem, paper proposes a Goal-Oriented Knowledge Copy network, GOKC. Specifically, goal-oriented discernment mechanism designed help model discern facts that highly correlated goal context. Besides, context manager devised copy not only discerned but also context, allows restate generated response. The empirical studies conducted two benchmarks results show our can significantly outperform several state-of-the-art models terms both automatic evaluation human judgments.

برای دانلود رایگان 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

Learning Spaces for Knowledge Generation

As the Internet is becoming the main point for information access, Libraries, Museums and similar Institutions are preserving their collections as digital object repositories. In that way, the important information associated with digital objects may be delivered as Internet content over portals equipped with modern interfaces and navigation features. This enables the virtualization of real inf...

متن کامل

Learning Spatial Knowledge for Text to 3D Scene Generation

We address the grounding of natural language to concrete spatial constraints, and inference of implicit pragmatics in 3D environments. We apply our approach to the task of text-to-3D scene generation. We present a representation for common sense spatial knowledge and an approach to extract it from 3D scene data. In text-to3D scene generation, a user provides as input natural language text from ...

متن کامل

Recursive Feature Generation for Knowledge-based Learning

When humans perform inductive learning, they often enhance the process with background knowledge. With the increasing availability of well-formed collaborative knowledge bases, the performance of learning algorithms could be significantly enhanced if a way were found to exploit these knowledge bases. In this work, we present a novel algorithm for injecting external knowledge into induction algo...

متن کامل

Optical klystron FEL scheme for x-ray coherent harmonic generation at ISLF storage ring

The possibility of inserting an optical klystron on the storage ring of an Iranian Light Source Facility as a Synchrotron Radiation Free electron Laser for coherence harmonic generation in the soft x-ray spectrum has been studied. The study has been performed using the 3D FEL simulation code GENESIS 1.3 and in accordance with the available characteristic parameters of ISLF. Different potential ...

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: Proceedings of the ... AAAI Conference on Artificial Intelligence

سال: 2021

ISSN: ['2159-5399', '2374-3468']

DOI: https://doi.org/10.1609/aaai.v35i14.17486