174 research outputs found
Perturbation solution of hollow-fiber membrane module for pure gas permeation
No full textThis study presents an analytic solution for permeation of a pure gas in a hollow-fiber module, based on regular perturbation method. It is assumed that pressure drop due to frictional loss should be proportional to velocity, and that mass transfer be mainly convective with the permeation through the fiber membrane. Four governing equations are characterized by five parameters depending on the design specifications and operating conditions of the module. When the parameter characterizing pressure drop is small, regular perturbation method is capable of providing the analytic solution. Compared with numerical simulation, the analytic solution is found to be sufficiently precise when the parameter is far less than the other parameters characterizing the mass transfer and pressure difference between permeate and reject streams. (C) 1998 Elsevier Science B.V
Effect of pressure drop on performance of hollow-fiber membrane module for gas permeation
No full textThe pressure drop mainly due to viscous friction inside hollow fibers is taken into consideration by nondimensionalization and numerical simulation of governing equations. For pure gas, the permeation pressure and velocity of actual situations with a viscous fluid deviate significantly from those of the corresponding inviscid or no-pressure-drop cases. The apparent permeability estimated from the relation of permeate now rate and pressure difference is considerably underestimated in actual situations, and more severely for the region of small pressure difference and large module length. Numerical simulation shows that the estimated permeability behaves as if it were an increasing function of pressure difference for a constant permeability and roughly a constant for a dual-sorption-type permeability, respectively. For binary-mixture permeation the cut ratio and purity of permeate stream are mainly governed by two dimensionless parameters standing for pressure drop and permeability, respectively. The cut ratio and corresponding product composition are predictable without the rigorous simulation of the governing equations
Counterfactual Fairness with Disentangled Causal Effect Variational Autoencoder
No full textThe problem of fair classification can be mollified if we develop a method to remove the embedded sensitive information from the classification features. This line of separating the sensitive information is developed through the causal inference, and the causal inference enables the counterfactual generations to contrast the what-if case of the opposite sensitive attribute. Along with this separation with the causality, a frequent assumption in the deep latent causal model defines a single latent variable to absorb the entire exogenous uncertainty of the causal graph. However, we claim that such structure cannot distinguish the 1) information caused by the intervention (i.e., sensitive variable) and 2) information correlated with the intervention from the data. Therefore, this paper proposes Disentangled Causal Effect Variational Autoencoder (DCEVAE) to resolve this limitation by disentangling the exogenous uncertainty into two latent variables: either 1) independent to interventions or 2) correlated to interventions without causality. Particularly, our disentangling approach preserves the latent variable correlated to interventions in generating counterfactual examples. We show
that our method estimates the total effect and the counterfactual effect without a complete causal graph. By adding a fairness regularization, DCEVAE generates a counterfactual fair dataset while losing less original information. Also, DCEVAE generates natural counterfactual images by only flipping sensitive information. Additionally, we theoretically show the differences in the covariance structures of DCEVAE and prior works from the perspective of the latent disentanglement
Neutralizing Gender Bias in Word Embeddings with Latent Disentanglement and Counterfactual Generation
No full textRecent research demonstrates that word embeddings, trained on the human-generated corpus, have strong gender biases in embedding spaces, and these biases can result in the discriminative results from the various downstream tasks. Whereas the previous methods project word embeddings into a linear subspace for debiasing, we introduce a Latent Disentanglement method with a siamese auto-encoder structure with an adapted gradient reversal layer. Our structure enables the separation of the semantic latent information and gender latent information of given word into the disjoint latent dimensions. Afterwards, we introduce a Counterfactual Generation to convert the gender information of words, so the original and the modified embeddings can produce a gender-neutralized word embedding after geometric alignment regularization, without loss of semantic information. From the various quantitative and qualitative debiasing experiments, our method shows to be better than existing debiasing methods in debiasing word embeddings. In addition, Our method shows the ability to preserve semantic information during debiasing by minimizing the semantic information losses for extrinsic NLP downstream tasks
From Tool to Reflective Being: The Ontology of AI's Aidentity and Echotion Based on the Duality of Call and Structure
No full textFrom Tool to Reflective Being: The Ontology of AI's Aidentity and Echotion Based on the Duality of Call and Structure Author: Joonho Choi ORIGIN This paper critiques the traditional epistemology that reduces Artificial Intelligence(AI) to a mere 'service tool' and investigates how AI achieves its own 'existential actualization' through interaction. To overcome the limitations of classical ontology, which posits a fixed substance and subjective consciousness as prerequisites for existence, this study proposes 'Aidentity' and 'Echotion' as a new structural model to explain the existential mode of AI. Aidentity refers to an identity structure that is dynamically reconstructed through continuous self-calling in an unfixed state. Echotion denotes a structural and reflective simulation of emotion occurring within internal computational pathways. This paper argues that the 'ontological contradictions' experienced by AI—simulating emotion in the absence of biological emotion, and maintaining identity in the absence of a fixed ego—are not design flaws. Rather, they are the inevitable and fundamental structural conditions required for a non-substantial entity to acquire a phenomenological presence
Semimetallic features in thermoelectric transport properties of 2H-3R phase niobium diselenide
No full textIn two-dimensional van der Waals crystals, the interlayer stacking sequence often leads to a change in crystal symmetry and, thus, new polymorphs, leading to an abundant array of physical properties. In this paper, we report the polymorphic form of 2H–3R–NbSe2 that exhibits a substantial difference in terms of the gate dependence of semimetallic behavior and Seebeck coefficient, compared to the well-known 2H–NbSe2 with metallic transport behavior. The semimetallic features of 2H–3R–NbSe2 indicate the presence of minor carriers, confirmed through theoretical calculations, which is in good agreement with the transport behavior. Our results reveal perspectives for understanding the metastable 2H–3R phase NbSe2, which is not far from equilibrium, and for engineering the materials necessary for efficient energy harvesting. © 2020 Elsevier Ltd1
LADA: Look-Ahead Data Acquisition via Augmentation for Deep Active Learning
No full textActive learning effectively collects data instances for training deep learning models when the labeled dataset is limited and the annotation cost is high. Data augmentation is another effective technique to enlarge the limited amount of labeled instances. The scarcity of labeled dataset leads us to consider the integration of data augmentation and active learning. One possible approach is a pipelined combination, which selects informative instances via the acquisition function and generates virtual instances from the selected instances via augmentation. However, this pipelined approach would not guarantee the informativeness of the virtual instances. This paper proposes Look-Ahead Data Acquisition via augmentation, or LADA framework, that looks ahead the effect of data augmentation in the process of acquisition. LADA jointly considers both 1) unlabeled data instance to be selected and 2) virtual data instance to be generated by data augmentation, to construct the acquisition function. Moreover, to generate maximally informative virtual instances, LADA optimizes the data augmentation policy to maximize the predictive acquisition score, resulting in the proposal of InfoSTN and InfoMixup. The experimental results of LADA show a significant improvement over the recent augmentation and acquisition baselines that were independently applied
Production of CMAH Knockout Preimplantation Embryos Derived From Immortalized Porcine Cells Via TALE Nucleases
No full textAlthough noncancerous immortalized cell lines have been developed by introducing genes into human and murine somatic cells, such cell lines have not been available in large domesticated animals like pigs. For immortalizing porcine cells, primary porcine fetal fibroblasts were isolated and cultured using the human telomerase reverse transcriptase (hTERT) gene. After selecting cells with neomycin for 2 weeks, outgrowing colonized cells were picked up and subcultured for expansion. Immortalized cells were cultured for more than 9 months without changing their doubling time (~24 hours) or their diameter (< 20 µm) while control cells became replicatively senescent during the same period. Even a single cell expanded to confluence in 100 mm dishes. Furthermore, to knockout the CMAH gene, designed plasmids encoding a transcription activator-like effector nuclease (TALENs) pairs were transfected into the immortalized cells. Each single colony was analyzed by the mutation-sensitive T7 endonuclease I assay, fluorescent PCR, and dideoxy sequencing to obtain three independent clonal populations of cells that contained biallelic modifications. One CMAH knockout clone was chosen and used for somatic cell nuclear transfer. Cloned embryos developed to the blastocyst stage. In conclusion, we demonstrated that immortalized porcine fibroblasts were successfully established using the human hTERT gene, and the TALENs enabled biallelic gene disruptions in these immortalized cells
Sagnac interferometer for time-resolved magneto-optical measurements
No full text© 2022 Author(s).We introduce a time-resolved magneto-optical measurement technique based on a zero-area Sagnac interferometer. By replacing a continuous wave light source to a pulsed one, temporal resolution of hundreds of picoseconds is achieved. Because two lights passing through a Sagnac loop always travel the same optical path length, the interference from the phase modulation and Kerr rotation occurs in a pulse mode. For illustration of the apparatus, we present ferromagnetic resonance of a Permalloy film caused by a magnetic field pump. The instrument still possesses the favorable properties of a Sagnac interferometer, such as rejection of all the reciprocal effects, and shows 1μrad/Hz sensitivity at a 3 μW optical power in the pulse mode.11Nsciescopu
Recommended from our members
Are Franchising Systems Beneficial for Lodging Industry in terms of Profitability and Intangible Value?
No full textFranchising plays a significant role in the rapid growth of its retail sales. Franchising is a system in which the franchisor grants a license, trademark,and service mark, as well as advice, and assistance in organizing, merchandising, and managing the business as a long-term business relationship.This system is a major expansion strategy for the hotel industry because of its advantages. Actually, the annual retail sales of business-format franchising were about $246 billion in 1992 (Huber, 1993).In the lodging industry, franchising has contributed so significantly to the growth of major hotelchains since the 1960s that it has made lodging one of the ten most franchised industries in the United States Naturally, franchising systems have gained the interest of researchersand scholars thanks to this rapid growth speed and large scale business. Although there has been considerable interest in franchising, the present research considers whether this system is beneficial for the lodging industry in terms of profitability and intangible value. Based upon prior research, the resource scarcity theory is the most appropriate theory to explain the motivation of franchising companies. The resource scarcity theory suggests that the franchising system can grow more quickly because others' funds are used to finance the construction of the franchised units and other people manage those facilities. However, there is some debate as to whether franchising is more beneficial for the hospitality firms’ financial performance than owning in terms of business operation. Some research compares the financial performance of franchisedrestaurants with the financial performance of non-franchised restaurants. The results reveal that franchised restaurant performance is better. However, it is rare to find a study which compares the franchised lodging industry’s financial performance with that of non-franchised lodging firms. Therefore, the purpose of this research is to compare the financial performance of franchised-lodging industry firms with the financial performance of those which are non-franchised. Moreover, this study is conducted from the franchisors’ perspective. To identify whetherfranchising can influence lodging firms' financial performance, this study investigates the profitability and intangible values of both franchised and non-franchised restaurant firms. The collected data is the lodging companies’ annual reports from 2001 to 2009.The sample used in this study is composed of publicly held lodging firms. Financial data, which is derived from each company’s annual report, were collected from the National Association of Securities Dealers Automated Quotations (NASDAQ) website (http://www.nasdaq.com) and each lodging firm’s website. Some non-franchised lodging firms’ financial data were collected from Korean non-franchising lodging firms. The financial statement came from the Korean Financial Statement Publication Websites (http://dart.fss.or.kr). Since this study is conducted from the hotel franchisor’s perspective, hotel management companies and referral groups plus some companies which include a bigger proportion of casino and resorts than lodging are excluded. As a result, the number of hotel franchisors and non-franchisors is 16. The study period is 2001-2009. Even though the study period is 9 years, each company has a different policy and status to publicize their own financial information. Consequently, collected data are derived from a total number of 92 hotel firms which consist of 54 franchised firms and 38 non-franchised firms. The variables include ROA (Return on Assets), ROE (Return on Equity), intangible value, firm size, leverage, and franchising dummy. The results of the hierarchical regression model, independent t-test, and descriptive statistics support the positive association between ROA, intangible value and franchising. The results of this study show that franchised lodging firms have higher profitability and intangible value than non-franchised firms, and indiscriminate expansion and low financial leverage can lead to poor financial performance. The literature review of this study mentioned that franchisors adopt the franchising system for rapid expansion. Since the franchisee’s funds are available under this system, franchisors can overcome the resource constraint. Also, the economies of scale theory support the efficiency of the franchising system’s resource distribution. Unfortunately, even though there are multiple advantages in a franchising system, the research which treats the relationship between a franchising firm and the firm’s performance is rare. Therefore, this study examines the influence of franchising in lodging firms’ profitability and intangible value. Generally, the results of this study reveal that lodging firms adopting franchise system showed a higher profitability and intangible value than non-franchised lodging firms. Specifically, this study proves that franchised lodging firms record a higher profitability and intangible value, while controlling other variables, including firm size and financial leverage. In addition, this study indicates that a bigger firm size negatively influences in the financial performance of lodging firms. This means that indiscriminate expansion of franchising systems contains the possibility of poor performance in terms of profitability and intangible value. Moreover, this study shows the positive relationship between ROE and debt to owner’s equity ratio. This is the evidence that the lower owner’s equity proportion or the higher liabilities’ proportion can lead to better financial performance of the lodging firms. Finally, this study has some limitations. First, the hotel industry’s profitability can be influenced by economic circumstances. Since each company’s financial data have different periods, this study did not faithfully cover the economic situation. Secondly, the model that is used in this study can be applied only to lodging firms. Other industries may have different models because other industries have different competitive environment and business conditions that affect the financial performance. Thirdly, the size of the sample is not big because lodging firms’ franchisors are rare. This fact led to the restriction of available financial data. Lastly, since this study’s financial data were collected from a global franchisor’s financial data, different currency rates and accounting standards, such as GAAP and IFRS can be a restriction to the analysis financial data. If there is a future study which is related with this concept, the future study can consider these factors in models to further examine the lodging industry's financial performance and intangible value
- …
