Supporting Biomimetic Design Through Categorization of Natural-Language Keyword-Search Results

2009 ◽  
Author(s):  
Ji Ke ◽  
J. S. Wallace ◽  
L. H. Shu
Keyword(s):  
Fuel Cell ◽  
Natural Language ◽  
Keyword Search ◽  
Bipolar Plate ◽  
Biological Information ◽  
Good Source ◽  
Search Results ◽  
Biological Phenomena ◽  
Keyword Searches

Biology is a good source of analogies for engineering design. One approach of retrieving biological analogies is to perform keyword searches on natural-language sources such as books, journals, etc. A challenge of retrieving information from natural-language sources is the potential requirement to process a large number of search results. This paper describes a categorization method that organizes a large group of diverse biological information into meaningful categories. The benefits of the categorization functionality are demonstrated through a case study on the redesign of a fuel cell bipolar plate. In this case study, our categorization method reduced the effort to systematically identify biological phenomena by up to ∼80%.

Supporting Biomimetic Design by Embedding Metadata in Natural-Language Corpora

2010 ◽  
Author(s):  
J. Ke ◽  
I. Chiu ◽  
J. S. Wallace ◽  
L. H. Shu
Keyword(s):  
Natural Language ◽  
Bipolar Plate ◽  
Word Sense ◽  
Biomimetic Design ◽  
Search Results ◽  
Part Of Speech ◽  
Biological Phenomena ◽  
Abstract Nouns ◽  
Keyword Searches

Biology is a good source of analogies for engineering design. One approach of retrieving biological analogies is to perform keyword searches on natural-language sources such as books, journals, etc. A challenge in retrieving information from natural-language sources is the potential requirement to process a large number of search results. This paper describes how inserting metadata such as part-of-speech, word sense and lexicographical data for each word in a natural-language source can help users identify relevant biological stimuli for biomimetic design. Although this research is still exploratory, initial qualitative observations demonstrate successful identification and separation of biological phenomena relevant to either desired functions or desired qualities. In addition, by incorporating the aforementioned metadata, we can automatically remove search results where search keywords act on abstract nouns or where keywords are used in irrelevant senses. The benefits of embedding metadata are demonstrated through a case study on the redesign of a fuel cell bipolar plate. In this case study, our method can be used to hide 64% of the search results that are unlikely to contain useful biological phenomena, reducing the effort to systematically identify relevant biological analogies.

scholarly journals Artificial intelligence-based conversational agent to support medication prescribing

JAMIA Open ◽  
2020 ◽  
Vol 3 (2) ◽  
pp. 225-232 ◽  
Author(s):  
Anita M Preininger ◽  
Brett South ◽  
Jeff Heiland ◽  
Adam Buchold ◽  
Mya Baca ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Natural Language ◽  
Subject Matter ◽  
System Architecture ◽  
Keyword Search ◽  
Conversational Agent ◽  
Subject Matter Experts ◽  
Rater Agreement ◽  
And Performance ◽  
Keyword Searches

Abstract Objective This article describes the system architecture, training, initial use, and performance of Watson Assistant (WA), an artificial intelligence-based conversational agent, accessible within Micromedex®. Materials and methods The number and frequency of intents (target of a user’s query) triggered in WA during its initial use were examined; intents triggered over 9 months were compared to the frequency of topics accessed via keyword search of Micromedex. Accuracy of WA intents assigned to 400 queries was compared to assignments by 2 independent subject matter experts (SMEs), with inter-rater reliability measured by Cohen’s kappa. Results In over 126 000 conversations with WA, intents most frequently triggered involved dosing (N = 30 239, 23.9%) and administration (N = 14 520, 11.5%). SMEs with substantial inter-rater agreement (kappa = 0.71) agreed with intent mapping in 247 of 400 queries (62%), including 16 queries related to content that WA and SMEs agreed was unavailable in WA. SMEs found 57 (14%) of 400 queries incorrectly mapped by WA; 112 (28%) queries unanswerable by WA included queries that were either ambiguous, contained unrecognized typographical errors, or addressed topics unavailable to WA. Of the queries answerable by WA (288), SMEs determined 231 (80%) were correctly linked to an intent. Discussion A conversational agent successfully linked most queries to intents in Micromedex. Ongoing system training seeks to widen the scope of WA and improve matching capabilities. Conclusion WA enabled Micromedex users to obtain answers to many medication-related questions using natural language, with the conversational agent facilitating mapping to a broader distribution of topics than standard keyword searches.

Bio-Inspired Solutions in Design for Manufacturing of Micro Fuel Cell

2014 ◽  
Author(s):  
Farzaneh Omidvarnia ◽  
Hans Nørgaard Hansen
Keyword(s):  
Fuel Cell ◽  
Direct Methanol Fuel Cell ◽  
Micro Manufacturing ◽  
Design Specifications ◽  
Biological Phenomena ◽  
Fuel Delivery ◽  
Direct Methanol ◽  
And Function ◽  
Methanol Fuel Cell

In this paper the application of biomimetic principles in design for micro manufacturing is investigated. A micro direct methanol fuel cell (μDMFC) for power generation in hearing aid devices is considered as the case study in which the bio-inspired functions are replicated. The focus in design of μDMFC is mainly on solving the problem of fuel delivery to the anode in the fuel chamber. Two different biological phenomena are suggested, and based on them different bio-inspired solutions are proposed and modeled in CAD software. Considering the manufacturing constraints and design specifications, the advantages and drawbacks of each proposed solution is discussed. Finally, the most feasible idea, in terms of manufacturing and function, is selected and an initial experimental verification is carried out.

Methods for Supporting Bioinspired Design

2011 ◽  
Author(s):  
Michael W. Glier ◽  
Joanna Tsenn ◽  
Julie S. Linsey ◽  
Daniel A. McAdams
Keyword(s):  
Ad Hoc ◽  
Biological Information ◽  
Functional Modeling ◽  
Graduate Degrees ◽  
Bioinspired Design ◽  
Powerful Approach ◽  
Engineering Problems ◽  
Keyword Searches ◽  
Design Inspiration

Bioinspired design, the use of nature to inspire solutions to engineering problems, is a powerful approach for innovation but is generally practiced on an ad hoc basis. Only recently, however, have researchers sought to develop formal tools and principles to effectively tap the wealth of design solutions found within nature. Research efforts in applying design tools such as the Theory of Inventive Problem Solving (TRIZ), functional modeling, and keyword searches to bioinspired design are summarized in this paper. The efforts to develop effective tools to search biological information for design inspiration are also discussed. This paper presents a case study where BioTRIZ, functional modeling, and bio-keyword searches were taught in a weekend-long workshop to working professionals in the process of obtaining their graduate degrees. The students were then given a project to re-engineer a product using the bioinspired methods. An analysis of their reports is given that discusses student success in implementing bioinspired design methods and student feedback on the methods.

Effects of Abstraction on Selecting Relevant Biological Phenomena for Biomimetic Design

2014 ◽  
Vol 136 (11) ◽  
Author(s):  
Tao Feng ◽  
Hyunmin Cheong ◽  
L. H. Shu
Keyword(s):  
Natural Language ◽  
Biological Knowledge ◽  
Design Problems ◽  
Biomimetic Design ◽  
Search Results ◽  
Biological Phenomena ◽  
Novice Designers

The natural-language approach to identifying biological analogies exploits the existing format of much biological knowledge, beyond databases created for biomimetic design. However, designers may need to select analogies from search results, during which biases may exist toward: specific words in descriptions of biological phenomena, familiar organisms and scales, and strategies that match preconceived solutions. Therefore, we conducted two experiments to study the effect of abstraction on overcoming these biases and selecting biological phenomena based on analogical similarities. Abstraction in our experiments involved replacing biological nouns with hypernyms. The first experiment asked novice designers to choose between a phenomenon suggesting a highly useful strategy for solving a given problem, and another suggesting a less-useful strategy, but featuring bias elements. The second experiment asked novice designers to evaluate the relevance of two biological phenomena that suggest similarly useful strategies to solve a given problem. Neither experiment demonstrated the anticipated benefits of abstraction. Instead, our abstraction led to: (1) participants associating nonabstracted words to design problems and (2) increased difficulty in understanding descriptions of biological phenomena. We recommend investigating other ways to implement abstraction when developing similar tools or techniques that aim to support biomimetic design.

Retrieving Causally Related Functions From Natural-Language Text for Biomimetic Design

2014 ◽  
Vol 136 (8) ◽  
Author(s):  
Hyunmin Cheong ◽  
L. H. Shu
Keyword(s):  
Natural Language ◽  
User Study ◽  
Keyword Search ◽  
Causal Relation ◽  
Biologically Inspired ◽  
Retrieval Method ◽  
Natural Language Text ◽  
Biological Phenomena ◽  
Biologically Inspired Design ◽  
Language Text

Identifying biological analogies is a significant challenge in biomimetic (biologically inspired) design. This paper builds on our previous work on finding biological phenomena in natural-language text. Specifically, a rule-based computational technique is used to identify biological analogies that contain causal relations. Causally related functions describe how one function is enabled by another function, and support the transfer of functional structure from analogies to design solutions. The causal-relation retrieval method uses patterns of syntactic information that represent causally related functions in individual sentences, and scored F-measures of 0.73–0.85. In a user study, novice designers found that of the total search matches, proportionally more of the matches obtained with the causal-relation retrieval method were relevant to design problems than those obtained with a single verb-keyword search. In addition, matches obtained with the causal-relation retrieval method increased the likelihood of using functional association to develop design concepts. Finally, the causal-relation retrieval method enables automatic extraction of biological analogies at the sentence level from a large amount of natural-language sources, which could support other approaches to biologically inspired design that require the identification of interesting biological phenomena.

Effects of Abstraction on Selecting Relevant Biological Phenomena for Biomimetic Design

2014 ◽  
Author(s):  
Tao Feng ◽  
Hyunmin Cheong ◽  
L. H. Shu
Keyword(s):  
Natural Language ◽  
Biological Knowledge ◽  
Design Problems ◽  
Biomimetic Design ◽  
Search Results ◽  
Biological Phenomena ◽  
Design By Analogy ◽  
Novice Designers

The natural-language approach to identifying biological analogies exploits the existing format of much biological knowledge, beyond databases created for biomimetic design. However, designers may need to select analogies from search results, during which biases may exist towards: specific words in descriptions of biological phenomena, familiar organisms and scales, and strategies that match preconceived solutions. Therefore, we conducted two experiments to study the effect of abstraction on overcoming these biases and selecting biological phenomena based on analogical similarities. Abstraction in our experiments involved replacing biological nouns with hypernyms. The first experiment asked novice designers to choose between a phenomenon suggesting a highly useful strategy for solving a given problem, and another suggesting a less-useful strategy, but featuring bias elements. The second experiment asked novice designers to evaluate the relevance of two biological phenomena that suggest similarly useful strategies to solve a given problem. Neither experiment demonstrated the anticipated benefits of abstraction. Instead, our abstraction led to: 1) novice designers associating non-abstracted words to design problems and 2) increased difficulty in understanding descriptions of biological phenomena. We recommend investigating other ways in which abstraction can be implemented when designing similar tools or techniques that aim to support biomimetic design and other design-by-analogy work.

Machine Learning of Motor Vehicle Accident Categories from Narrative Data

1996 ◽  
Vol 35 (04/05) ◽  
pp. 309-316 ◽  
Author(s):  
M. R. Lehto ◽  
G. S. Sorock
Keyword(s):  
Machine Learning ◽  
Bayesian Model ◽  
Keyword Search ◽  
Motor Vehicle ◽  
Motor Vehicle Accident ◽  
Computer Search ◽  
Vehicle Accident ◽  
Learning Technique ◽  
Expert Ratings ◽  
Keyword Searches

Abstract:Bayesian inferencing as a machine learning technique was evaluated for identifying pre-crash activity and crash type from accident narratives describing 3,686 motor vehicle crashes. It was hypothesized that a Bayesian model could learn from a computer search for 63 keywords related to accident categories. Learning was described in terms of the ability to accurately classify previously unclassifiable narratives not containing the original keywords. When narratives contained keywords, the results obtained using both the Bayesian model and keyword search corresponded closely to expert ratings (P(detection)≥0.9, and P(false positive)≤0.05). For narratives not containing keywords, when the threshold used by the Bayesian model was varied between p>0.5 and p>0.9, the overall probability of detecting a category assigned by the expert varied between 67% and 12%. False positives correspondingly varied between 32% and 3%. These latter results demonstrated that the Bayesian system learned from the results of the keyword searches.

Sign in / Sign up

Export Citation Format

Share Document