Automatic Discovery of Design Task Structure Using Deep Belief Nets

Abstract

With the arrival of cyber physical world and an extensive support of advanced information technology (IT) infrastructure, nowadays it is possible to obtain the footprints of design activities through emails, design journals, change logs, and different forms of social data. In order to manage a more effective design process, it is essential to learn from the past by utilizing these valuable sources and understand, for example, what design tasks are actually carried out, their interactions, and how they impact each other. In this paper, a computational approach based on the deep belief nets (DBN) is proposed to automatically uncover design tasks and quantify their interactions from design document archives. First, a DBN topic model with real-valued units is developed to learn a set of intrinsic topic features from a simple word-frequency-based input representation. The trained DBN model is then utilized to discover design tasks by unfolding hidden units by sets of strongly connected words, followed by estimating the interactions among tasks on the basis of their co-occurrence frequency in a hidden topic space. Finally, the proposed approach is demonstrated through a real-life case study using a design email archive spanning for more than 2 yr.

Process bottlenecks identification and its root cause analysis using fusion-based clustering and knowledge graph

“…To solve this limitation, advanced process information extraction approaches were conducted. For example, Lijun Lan et al focused on design process knowledge extraction and design process discovery from email data without explicit process information [9]. She proposed a deep belief net (DBN) that can automatically extract hidden process information and then model the design process.…”

“…Another reason is the difference between the manufacturing process and the knowledgeintensive process. The former is formal and repetitive, while the latter is flexible and unpredictable [9]. In the manufacturing process, the bottleneck can be identified from a local perspective that observes a quantitative metric of a single object, usually the machine [10].…”

Natural language processing in-and-for design research

“…Wasiak et al (2010, p. 58) analyse emails to discover topics such as functions, performance, features, operating environment, materials, manufacturing, cost and ergonomics. From the email exchanges in a traffic wave project, Lan, Liu, and Lu (2018, p. 7) map wordfrequency vectors and topic vectors (tasks, timestamps, persons, organisations, locations, input/output, techniques/tools) using Deep Belief Network -DBN (Bengio 2009;Lan, Liu, and Feng Lu 2017). Goepp et al (2019, p. 165) identify the following speech acts from email exchanges: Information, Explication, Evaluation, Description and Request.…”

A Data-Driven Text Mining and Semantic Network Analysis for Design Information Retrieval

“…As discussed above, one of the most commonly used tools for D 3 is text mining as well as natural language processing, since most of the electronic and digital data are texts in nature (Ur-Rahman and Harding, 2012). Lan et al (2017) proposed an approach based on deep belief net (DBN) to automatically discover design tasks and quantify their interactions from design email archive, where the deep belief neural network is used to learn a set of latent topic features from a simple word-frequency based input representation of the document. Dong and Agogino (1997) used natural language processing to induce a representation of design based on the analysis of syntactic pattern contained in the corpus of design documents.…”

“…The use of clustering on ECR can significantly help summarise the main features and changes added from the product development process of previous projects. Similarly, topic modelling has been implemented to analyse the past design emails archive in order to uncover design tasks and quantify their interaction (Lan et al, 2017). This helps the designers learn for the past, for example, what design tasks are actually carried out, and how they impact each other.…”