Data Driven Techniques for Organizing Scientific Articles Relevant to Biomimicry

Zhao, Yuanshuo; Baldini, Ioana; Sattigeri, Prasanna; Padhi, Inkit; Lee, Yoong Keok; Smith, Ethan

doi:10.1145/3278721.3278755

Cited by 3 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, the similarities between papers can be easily computed based on their BoW representations by considering their overlapping term features. In (Zhao et al, 2018), the authors have also investigated the use of Glove word embedding models (Pennington, Socher, & Manning, 2014) for generating the feature representations of papers in the biomimicry domain based on their abstract words.…”

Section: Supervised Approachesmentioning

confidence: 99%

“…They require the use of only labelled training data for classifier training. Some studies (Łukasik, Kuśmierczyk, Bolikowski, & Nguyen, 2013; Roul & Sahoo, 2017; Santos & Rodrigues, 2009; Surkis et al, 2016; Vogrinčič & Bosnić, 2011; Zhao et al, 2018) use text‐based techniques for mining papers' content information, which can be extracted from their different sections (i.e., abstract, title, keywords and full texts). Their main assumption is that papers conducted in the same research fields, use approximately the same terms/keywords for describing the papers research goals, backgrounds and conclusions.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A co‐training‐based approach for the hierarchical multi‐label classification of research papers

et al. 2020

View full text Add to dashboard Cite

This paper focuses on the problem of the hierarchical multi-label classification of research papers, which is the task of assigning the set of relevant labels for a paper from a hierarchy, using reduced amounts of labelled training data. Specifically, we study leveraging unlabelled data, which are usually plentiful and easy to collect, in addition to the few available labelled ones in a semi-supervised learning framework for achieving better performance results. Thus, in this paper, we propose a semisupervised approach for the hierarchical multi-label classification task of research papers based on the well-known Co-training algorithm, which exploit content and bibliographic coupling information as two distinct papers' views. In our approach, two hierarchical multi-label classifiers, are learnt on different views of the labelled data, and iteratively select their most confident unlabelled samples, which are further added to the labelled set. The success of our suggested Co-training-based approach lies in two main components. The first is the use of two suggested selection criteria (i.e., Maximum Agreement and Labels Cardinality Consistency) that enforce selecting confident unlabelled samples. The second is the appliance of an oversampling method that rebalances the labels distribution of the initial labelled set, which reduces the reinforcement of the label imbalance issue during the Co-training learning. The proposed approach is evaluated using a collection of scientific papers extracted from the ACM digital library. Performed experiments show the effectiveness of our approach with regards to several baseline methods.

show abstract

Section: Supervised Approachesmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

A co‐training‐based approach for the hierarchical multi‐label classification of research papers

et al. 2020

View full text Add to dashboard Cite

show abstract

“…More recently, Zhao et al demonstrated how crowdsourcing annotations of scientific articles containing biomimicryrelated content can facilitate a supervised learning approach to classifying scientific articles (Zhao et al 2018;Vattam and Goel 2011b). However, generating labels for a large dataset can again be expensive and difficult.…”

Section: Introductionmentioning

confidence: 99%

Imitation of Life: A Search Engine for Biologically Inspired Design

Emuna,

Borenstein,

Qian

et al. 2024

AAAI

View full text Add to dashboard Cite

Biologically Inspired Design (BID), or Biomimicry, is a problem-solving methodology that applies analogies from nature to solve engineering challenges. For example, Speedo engineers designed swimsuits based on shark skin. Finding relevant biological solutions for real-world problems poses significant challenges, both due to the limited biological knowledge engineers and designers typically possess and to the limited BID resources. Existing BID datasets are hand-curated and small, and scaling them up requires costly human annotations. In this paper, we introduce BARcode (Biological Analogy Retriever), a search engine for automatically mining bio-inspirations from the web at scale. Using advances in natural language understanding and data programming, BARcode identifies potential inspirations for engineering challenges. Our experiments demonstrate that BARcode can retrieve inspirations that are valuable to engineers and designers tackling real-world problems, as well as recover famous historical BID examples. We release data and code; we view BARcode as a step towards addressing the challenges that have historically hindered the practical application of BID to engineering innovation.

show abstract

Combining BioTRIZ and Multi-Factor Coupling for Bionic Mechatronic System Design

Wang,

2024

Applied Sciences

View full text Add to dashboard Cite

To realize the design process of bionic mechatronic systems, involving mapping from engineering to biology and inversion from biology to engineering, a novel design paradigm is introduced that integrates BioTRIZ with multi-factor coupling bionics. In the mapping stage from engineering to biology, BioTRIZ is employed to frame the concrete engineering issue as a general conflicting problem. The biological solution is refined by amalgamating the BioTRIZ solution derived from the contradiction matrix with biological instances. In the inversion stage of biology to engineering, a novel approach is proposed for constructing a bionic multi-factor coupling model, drawing inspiration from the establishment of biological multi-factor coupling model. This allows for a seamless correspondence between biological elements, such as morphology and behavior, and their respective engineering counterparts, including structure and algorithms. This correspondence ultimately achieves the engineering conceptual model that is rooted in biological principles. The practical application of this methodology is exemplified through a multi-biometric fusion bionic active vision system, underscoring its feasibility and efficacy.

show abstract

Data Driven Techniques for Organizing Scientific Articles Relevant to Biomimicry

Cited by 3 publications

References 26 publications

A co‐training‐based approach for the hierarchical multi‐label classification of research papers

A co‐training‐based approach for the hierarchical multi‐label classification of research papers

Imitation of Life: A Search Engine for Biologically Inspired Design

Combining BioTRIZ and Multi-Factor Coupling for Bionic Mechatronic System Design

Contact Info

Product

Resources

About