Deeply Curious Computational Design

نویسندگان

  • Kazjon Grace
  • Mary Lou Maher
چکیده

This paper describes ongoing work into a model of computational design that is based on curiosity. The model uses goal reformulation to pursue specific unexpected stimuli in order to generate new designs that are creative (i.e. novel and valuable). The model combines a number of techniques from AI, including goal reasoning, heuristic search, genetic algorithms and case-based reasoning, with a deep unsupervised neural network model of expectation. The deep learning model is used by the system to develop a representation of the space of possible designs that respects its latent structure, and can be used as the basis for goal-driven curious search. From this hidden-layer representation similarity functions are derived as well as evaluations of the surprisingness of particular combinations of features. In this position paper we summarize our previous contributions and outline our planned future research in the area of deep learning for AI in design.

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

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A CURIOUS DESIGN AGENT A Computational Model of Novelty-Seeking Behaviour in Design

This paper presents a “curious design agent”, i.e. an agent that uses the search for novel designs to guide its design actions. A computational model of curiosity based on a process called novelty detection is presented. The behaviour of the computational model is illustrated with a curious design agent searching the space of twodimensional patterns generated by a simulated Spirograph is report...

متن کامل

Designing for Interest and Novelty Motivating Design Agents

This paper is concerned with the motivation of design agents to promote the exploration of design spaces. A general form of motivation common to designers is a curiosity to discover interesting designs. This paper presents computational models of interest and curiosity based on the detection of novelty. We illustrate the behaviour of our model of interest in emergent properties of a design and ...

متن کامل

A Curious Agent for Network Anomaly Detection

This paper presents a novel approach to intrusion detection using curious agents to detect anomalies in network data. Curious agents use computational models of noveltyseeking behavior and interest, based on human curiosity, to reason about their experiences in their environment. They are online, single-pass agents that respond to the similarity, frequency and recentness of their experiences. A...

متن کامل

A curious agent for network anomaly detection

This paper presents a novel approach to intrusion detection using curious agents to detect anomalies in network data. Curious agents use computational models of noveltyseeking behavior and interest, based on human curiosity, to reason about their experiences in their environment. They are online, single-pass agents that respond to the similarity, frequency and recentness of their experiences. A...

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

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2016