The online shopping has been much easier and popular, and meanwhile brings new challenges and opportunities to the field of product design and marketing sale. On one hand, product manufacturers find it challenging to produce new popularly accepted products to meet the customers’ needs; on the other hand, end customers usually feel it difficult to buy ideal goods that they really want, even if navigating a huge amount of commodities. There are indeed a ‘communication gap’ between the customers and manufacturers. As an effort to partially resolve the issue, this paper proposes a novel product synthesis approach from ‘voice of the customer’ over product knowledge graphs. Here the voice of customers mainly refer to the buyers’ product reviews from online shopping platforms or blogs, while the product knowledge graph is constructed containing professional hierarchical product knowledge on its properties based on ontological models. Using the technologies of natural language processing, we first extract the customs’ polarities on each specific aspect of a product, which are then transited to design requirements on the product’s design components. Based on the requirement extractions, and the pre-built product knowledge, semantic web and reasoning techniques are utilized to synthesize a novel product that meets more customer needs. Typical case studies on mobile phones from raw online data demonstrate the proposed approach’s performance.

The descriptions of capabilities of manufacturing companies can be found in multiple locations including company websites, legacy system databases, and ad hoc documents and spreadsheets. The capability descriptions are often represented using natural language. To unlock the value of unstructured capability information and learn from it, there is a need for developing advanced quantitative methods supported by machine learning and natural language processing techniques. This research proposes a multi-step unsupervised learning methodology using K-means clustering and topic modeling techniques in order to build clusters of suppliers based on their capabilities, extract and organize the manufacturing capability terminology, and discover nontrivial patterns in manufacturing capability corpora. The capability data is extracted either directly from the website of manufacturing firms or from their profiles in e-sourcing portals and directories. Feature extraction and dimensionality reduction process in this work in supported by Ngram extraction and Latent Semantic Analysis (LSA) methods. The proposed clustering method is validated experimentally based a dataset composed of 150 capability descriptions collected from web-based sourcing directories such as the Thomas Net directory for manufacturing companies. The results of the experiment show that the proposed method creates supplier cluster with high accuracy.

This article describes the use of machine learning and Natural Language Processing to determine the quality of existing technical instructions. Technical instructions were gathered from publicly available, online sources and analyzed to determine their usability and adherence level to published guidelines in the engineering and in the technical communication fields. Although some organizations have established guidelines unique to their industries, analysis of the randomly-selected instruction sets shows that documents often fall short of targets associated with usability, clarity, and benefit to the audience. The performed analyses show that over fifty percent of the instructions lack compliance with at least one established guideline.

With the advent of the big-data era, massive textual information stored in electronic and digital documents have become valuable resources for knowledge discovery in the fields of design and engineering. Ontology technologies and semantic networks have been widely applied with text mining techniques including Natural Language Processing (NLP) to extract structured knowledge associations from the large-scale unstructured textual data. However, most existing works mainly focus on how to construct the semantic networks by developing various text mining methods such as statistical approaches and semantic approaches, while few studies are found to focus on how to subsequently analyze and fully utilize the already well-established semantic networks. In this paper, a specific network analysis method is proposed to discover the implicit knowledge associations from the existing semantic network for improving knowledge discovery and design innovation. Pythagorean means are applied with Dijkstra’s shortest path algorithm to discover the implicit knowledge associations either around a single knowledge concept or between two concepts. Six criteria are established to evaluate and rank the correlation degree of the implicit associations. Two engineering case studies were conducted to illustrate the proposed knowledge discovery process, and the results showed the effectiveness of the retrieved implicit knowledge associations on helping providing relevant knowledge from various aspects, and provoking creative ideas for engineering innovation.

Open innovation often enjoys large quantities of submitted content. Yet the need to effectively process such large quantities of content impede the widespread use of open innovation in practice. This article presents an exploration of needs-based open innovation using state-of-the art natural language processing (NLP) algorithms to address existing limitations of exploiting large amounts of incoming data. The Semantic Textual Similarity (STS) algorithms were specifically developed to compare sentence-length text passages and were used to rate the semantic similarity of pairs of text sentences submitted by users of a custom open innovation platform. A total of 341 unique users submitted 1,735 textual problem statements or unmet needs relating to multiple topics: cooking, cleaning, and travel. Scores of equivalence generated by a consensus of ten human evaluators for a subset of the needs provided a benchmark for similarity comparison. The semantic analysis allowed for rapid (1 day per topic), automated screening of redundancy to facilitate identification of quality submissions. In addition, a series of permutation analyses provided critical crowd characteristics for the rates of redundant entries as crowd size increases. The results identify top modern STS algorithms for needfinding. These predicted similarity with Pearson correlations of up to .85 when trained using need-based training data and up to .83 when trained using generalized data. Rates of duplication varied with crowd size and may be approximately linear or appear asymptotic depending on the degree of similarity used as a cutoff. Semantic algorithm performance has shown rapid improvements in recent years. Potential applications to screen duplicates and also to screen highly unique sentences for rapid exploration of a space are discussed.

The objective of this research is to investigate the requirements and performance of parts-of-speech tagging of assembly work instructions. Natural Language Processing of assembly work instructions is required to perform data mining with the objective of knowledge reuse. Assembly work instructions are key process engineering elements that allow for predictable assembly quality of products and predictable assembly lead times. Authoring of assembly work instructions is a subjective process. It has been observed that most assembly work instructions are not grammatically complete sentences. It is hypothesized that this can lead to false parts-of-speech tagging (by Natural Language Processing tools). To test this hypothesis, two parts-of-speech taggers are used to tag 500 assembly work instructions (obtained from the automotive industry). The first parts-of-speech tagger is obtained from Natural Language Processing Toolkit (nltk.org) and the second parts-of-speech tagger is obtained from Stanford Natural Language Processing Group (nlp.stanford.edu). For each of these taggers, two experiments are conducted. In the first experiment, the assembly work instructions are input to the each tagger in raw form. In the second experiment, the assembly work instructions are preprocessed to make them grammatically complete, and then input to the tagger. It is found that the Stanford Natural Language Processing tagger with the preprocessed assembly work instructions produced the least number of false parts-of-speech tags.

Formal ontology and rule-based approaches founded on semantic technologies have been proposed as powerful mechanisms to enable early manufacturability feedback. A fundamental unresolved problem in this context is that all manufacturing knowledge is encoded in unstructured text and there are no reliable methods to automatically convert it to formal ontologies and rules. It is impractical for engineers to write accurate domain rules in a structured semantic languages such as Web Ontology Language (OWL) or Semantic Application Design Language (SADL). Previous efforts in manufacturing research that have targeted extraction of OWL ontologies from text have focused on basic concept names and hierarchies. This paper presents a semantics-based framework for acquiring more complex manufacturing knowledge, primarily rules, in a semantically-usable form from unstructured English text such as those written in manufacturing handbooks. The approach starts with existing domain knowledge in the form of OWL ontologies and applies natural language processing techniques to extract dependencies between different words in the text that contains the rule. Domain-specific triples capturing each rule are then extracted from each dependency graph. Finally, new computer-interpretable rules are composed from the triples. The feasibility of the framework has been evaluated by automatically and accurately generating rules for manufacturability from a manufacturing handbook. The paper also documents the cases that result in ambiguous results. Analysis of the results shows that the proposed framework can be extended to extract domain ontologies which forms part of the ongoing work that also focuses on addressing challenges to automate different steps and improve the reliability of the system.

We examined online product reviews as a source of novel affordances. Certain affordances may only be discovered through extended use across various environments. User-generated reviews may thus contain unique insights. We analyzed online consumer product reviews from Canadian Tire, one of Canada’s largest retailers. We determined properties of this collection of reviews and commonalities between valuable reviews. In addition to typical challenges associated with natural-language processing, e.g. word-sense disambiguation, we identify characteristics of online consumer reviews that create additional challenges. These challenges include the use of ‘wild English’ and sarcasm in online reviews. We first present criteria to define and more objectively identify novel affordances from review content. Next, k-means clustering reveals that a combination of syntactical features and high frequency word percentages can separate descriptive from non-descriptive review content. Finally, we identified cue phrases that may indicate higher likelihood of affordance content in a review. Despite existing obstacles, the substantial volume of available online product reviews has potential to become a valuable source of affordances and feedback for designers and retailers alike.

The United States generates more than 250 million tons of municipal solid waste (trash/garbage), with only 34% being recycled. In the broader global environment, the problem of waste management is becoming increasingly relevant, demanding innovative solutions. Traditional End-of-Life (EOL) approaches to managing waste include recycle, reuse, remanufacture and disposal. Recently, resynthesis was proposed as an alternative to traditional EOL options that combines multiple products to create a new product distinct from its parent assemblies. Resynthesis employs data mining and natural language processing algorithms to quantify assembly/subassembly combinations suitable for new product combinations. However, existing resynthesis methodologies proposed in the design community have been limited to exploring subassembly combinations, failing to explore potential combinations on a materials level. The authors of this paper propose a material resynthesis methodology that combines the materials of multiple EOL products using conventional manufacturing processes that generate candidate resynthesized materials that satisfy the needs of existing domains/applications. Appropriate applications for a resynthesized material are discovered by comparing the properties of the new material to the functional requirements of application classes which are found using clustering and latent semantic analysis. In the course of this paper, the authors present a case study that demonstrates the feasibility of the proposed material resynthesis methodology in the construction materials domain.

The objective of this research is to mine textual data (e.g., online discussion forums) generated by students enrolled in Massive Open Online Courses (MOOCs) in order to quantify students’ sentiment, in relation to their course performance. Massive Open Online Courses (MOOCs) are free to anyone with a computing device and a means of connecting to the internet and serve as a new paradigm for distance based education. While student interactions in traditional based brick and mortar classes are readily observable by students and instructors, quantifying the sentiments expressed by students in MOOCs remains challenging. This is in part due to the quantity of textual data being generated by students enrolled in MOOCs, in addition to a lack of quantitative methodologies that discover latent, previously unknown knowledge pertaining to student interactions and sentiments in the digital world. The authors of this work introduce a data mining driven methodology that employs natural language processing techniques and text mining algorithms to quantify students’ sentiments, based on their textual data provided during course assignment discussions. The researchers of this work aim to help educators understand the factors that may impact student performance, team interactions and overall learning outcomes in digital environments such as MOOCs.

Patents contain valuable information for engineering design. However, the increasing number of annual patent publications makes it difficult for any individual designer to assimilate all up-to-date knowledge hidden in patent documents. In this paper, we proposed a computational approach to interpret design structure embedded in patent claims using pre-developed ontology libraries. The study combined natural language processing (NLP) techniques, text data-mining, ontological engineering, and our rule-based tree generation method. Data sources and adopted tools included online patent documents, knowledge gathered from engineering textbooks, WordNet, a part-of-speech tagger developed by the Stanford NLP group, and Graphviz. We showed that the framework proposed in the paper not only could help minimize manual work required for obtaining design structures but also enable automatic dissimilarity comparison between patents.

Identifying relevant analogies from biology is a significant challenge in biomimetic design. Our natural-language approach addresses this challenge by developing techniques to search biological information in natural-language format, such as books or papers. This paper presents the application of natural-language processing techniques, such as part-of-speech tags, typed-dependency parsing, and syntactic patterns, to automatically extract and categorize causally related functions from text with biological information. Causally related functions, which specify how one action is enabled by another action, are considered important for both knowledge representation used to model biological information and analogical transfer of biological information performed by designers. An extraction algorithm was developed and scored F-measures of 0.78–0.85 in an initial development test. Because this research approach uses inexpensive and domain-independent techniques, the extraction algorithm has the potential to automatically identify patterns of causally related functions from a large amount of text that contains either biological or design information.

In the design process, the requirements serve as the benchmark for the entire product. Therefore, the quality of requirement statements is essential to the success of a design. Because of their ergonomic-nature, most requirements are written in natural language (NL). However, writing requirements in natural language presents many issues such as ambiguity, specification issues, and incompleteness. Therefore, identifying issues in requirements involves analyzing these NL statements. This paper presents a linguistic approach to requirement analysis, which utilizes grammatical elements of requirements statements to identify requirement statement issues. These issues are organized by the entity—word, sentence, or document—that they affect. The field of natural language processing (NLP) provides a core set of tools that can aid with this linguistic analysis and provide a method to create a requirement analysis support tool. NLP addresses requirements on processing levels: lexical, syntactic, semantic, and pragmatic. While processing on the lexical and syntactic level are well-defined, mining semantic and pragmatic data is performed in a number of different methods. This paper provides an overview of these current requirement analysis methods in light of the presented linguistic approach. This overview will be used to identify areas for further research and development. Finally, a prototype requirement analysis support tool will be presented. This tool seeks to demonstrate how the semantic processing level can begin to be addressed in requirement analysis. The tool will analyze a sample set of requirements from a family of military tactical vehicles (FMTV) requirements document. It implements NLP tools to semantically compare requirements statements based upon their grammatical subject.

Robotic alternative to many manual operations falls short in application due to the difficulties in capturing the manual skill of an expert operator. One of the main problems to be solved if robots are to become flexible enough for various manufacturing needs is that of end-user programming. An end-user with little or no technical expertise in robotics area needs to be able to efficiently communicate its manufacturing task to the robot. This paper proposes a new method for robot task planning using some concepts of Artificial Intelligence. Our method is based on a hierarchical knowledge representation and propositional logic, which allows an expert user to incrementally integrate process and geometric parameters with the robot commands. The objective is to provide an intelligent and programmable agent such as a robot with a knowledge base about the attributes of human behaviors in order to facilitate the commanding process. The focus of this work is on robot programming for manufacturing applications. Industrial manipulators work with low level programming languages. This work presents a new method based on Natural Language Processing (NLP) that allows a user to generate robot programs using natural language lexicon and task information. This will enable a manufacturing operator (for example for painting) who may be unfamiliar with robot programming to easily employ the agent for the manufacturing tasks.

In engineering design, establishing functional requirements (FRs) based on customer needs (CNs) is critical to ensure the success of a product. However, it may take a lot of time for a designer to state functional requirements explicitly. To overcome this difficulty, the authors have proposed a method to interpret design intention in accordance with the axiomatic design (AD) framework by applying natural language processing (NLP) techniques on patents. The goal of the proposed method is to provide functional requirements and design parameters that are automatically extracted for designers. Although the initial result shows that the method is dependent on the writing style used in a patent, this fresh attempt demonstrates the feasibility of automatically interpreting design intention.

Conceptual design is a process wherein new functions are created through engineering design. In conceptual design, we use natural language since it plays an important role in the expression and operation of a function. Moreover, natural language is used in our day-to-day thinking processes and is expected to keep a fine interface with the designer. However, it is at a disadvantage with regard to the expression of a function, since physical phenomena, which are the essence of a function, are better expressed in the form of mathematical equations than natural languages. In this study, we attempt to develop a method for using natural language for operating a function by harnessing its advantages and overcoming its disadvantage. We focus on the vital process in conceptual design, that is, the function dividing process wherein the required function is decomposed into sub functions that satisfy the required function. We construct a thesaurus by semiautomatic extraction of the hierarchical structures of words from a document by using natural language processing. We show that the constructed thesaurus can be useful in supporting the function dividing process.

Usability and versatility are two of the most important issues when using modern 3D simulation systems within the field of automation technology and virtual environments. 3D simulations and virtual worlds proved to be versatile tools to program, supervise and command complex robotic and automation systems. For industrial robots, 3D simulation systems like COSIMIR ® introduced the so called Native Language Programming (NLP) concept enabling the automation expert to program each robot using its native programming language. But what about programming other automation components or other dynamic components in virtual environments, what about user friendly, intuitive graphical programming languages, what about easy-to-use worker oriented programming languages? When talking about graphical programming languages to model dynamic behavior, questions like “which graphical modeling languages should be supported?”, “which are the most powerful ones?” and “which one matches the most to my concrete application?” have to be answered. Each graphical programming language has its own advantages and disadvantages, so that the answer to all these questions has to be: Offer a choice of graphical modeling languages to the user and leave the decision to him. The advantage of this strategy is obvious: Instead of learning how to use a concrete modeling language or worrying about programming details, the user can focus on his individual automation task and so quickly build efficient solutions. Therefore this paper extends the NLP approach to graphical programming languages using a new kind of object oriented Petri Nets as an intermediate language. This enables the user to use — at the same time — finite automata like mealy machines or extended automata, activity diagrams as defined in UML 2, flowchart like diagrams (e. g. icon-based programming) and many more to model the dynamics or the behavior of dynamic components.

Textual documents are the most common way of storing and distributing information within organizations. Extracting useful information from large text collections is therefore the goal of every organization that would like to take advantage of the experience encapsulated in those texts. Entering data using a free text style is easy, as it does not require any special training. However, unstructured texts pose a major challenge for automatic extraction and retrieval systems. Generally, deep levels of text analysis using advanced and complex linguistic processing are necessary that involve computational linguistic experts and domain experts. Linguistic experts are rare in engineering organizations, which thus find it difficult to apply and exploit such advanced extraction techniques. It is therefore desirable to minimize the extensive involvement of linguist experts by learning extraction patterns automatically from example texts. In doing so, the analysis of given texts is necessary in order to identify the scope and suitable automatic methods. Focusing on causality reasoning in the field of fault diagnosis, the results of experimenting with an automatic causality extraction method using shallow linguistic processing are presented.