| 1. |
|
|
| 2. |
- Berglund, Per, et al.
(författare)
-
2-Methylalkanoic acids resolved by esterification catalysed by lipase from Candida rugosa : Alcohol chain length and enantioselectivity
- 1993
-
Ingår i: Tetrahedron Asymmetry. - 0957-4166. ; 4:8, s. 1869-1878
-
Tidskriftsartikel (refereegranskat)abstract
- Enantiomerically pure (R)-2-methyldecanoic acid and (S)-2-methyl-1-decanol were prepared in a multi gram scale by esterification reactions catalysed by lipase from Candida rugosa. The enantiomeric ratios (E-values) were determined as a function of the chain length of the alcohol used as the complementary substrate in cyclohexane. In the resolution of 2-methyldecanoic acid the highest value (E = 37 ± 5) was obtained, when either 2-hexanol, 1-heptanol or 1-octanol were used. In contrast, when resolving 2-methyloctanoic acid, the E-values increased continually with increasing chain length of the alcohol used. 1-Hexadecanol gave the highest value: E > 100. The E-values were determined from the enantiomeric excess (ee) of the product at a conversion below 0.4. After two consecutive esterification reactions enantiomerically pure (R)-2-methyldecanoic acid, >99.8% ee, and after subsequent reduction of the ester produced, (S)-2-methyl-1-decanol, 96.7% ee, were obtained.
|
|
| 3. |
- Berglund, Per, et al.
(författare)
-
Reversed enantiopreference of Candida rugosa lipase supports different modes of binding enantiomers of a chiral acyl donor
- 1998
-
Ingår i: Journal of Molecular Catalysis - B Enzymatic. - 1381-1177 .- 1873-3158. ; 5:1-4, s. 283-287
-
Tidskriftsartikel (refereegranskat)abstract
- Molecular modelling identifies two different productive modes of binding the enantiomers of a 2-methyldecanoic acid ester to the active site of Candida rugosa lipase (CRL). The fast reacting S-enantiomer occupies the previously identified acyl-binding tunnel of the enzyme, whereas the R- enantiomer leaves the tunnel empty. The modelling suggested that if both enantiomers were forced to bind to the active site leaving the tunnel empty, the enzyme would reverse its enantiopreference to become R-enantioselective. To test this hypothesis, we designed a structural analogue to 2- methyldecanoic acid, 2-methyl-6-(2-thienyl)hexanoic acid, which was expected to be too bulky to fit its acyl moiety into the acyl-binding tunnel. The CRL- catalysed hydrolysis of the ethyl ester of this substrate resulted in the preferential conversion of the R-enantiomer as predicted by molecular modelling. This represents the first kinetic evidence supporting the existence of two different modes of binding the enantiomers of a 2- methyldecanoic acid ester to the active site of CRL. We have shown that a rational 3D based approach in combination with substrate engineering can be used to predict and control the stereochemical outcome of a lipase catalysed reaction.
|
|
| 4. |
- Berglund, Per, et al.
(författare)
-
Switched enantiopreference of Humicola lipase for 2-phenoxyalkanoic acid ester homologs can be rationalized by different substrate binding modes
- 1999
-
Ingår i: Tetrahedron. - 0957-4166 .- 1362-511X. ; 10:21, s. 4191-4202
-
Tidskriftsartikel (refereegranskat)abstract
- Humicola lanuginosa lipase was used for enantioselective hydrolyses of a series of homologous 2-phenoxyalkanoic acid ethyl esters. The enantioselectivity (E-value) of the enzyme changed from an (R)-enantiomer preference for the smallest substrate, 2-phenoxypropanoic acid ester, to an (S)-enantiomer preference for the homologous esters with longer acyl moieties. The E-values span the range from E=13 (R) to E=56 (S). A molecular modeling study identified two different substrate-binding modes for each enantiomer. We found that the enantiomers favored different modes. This discovery provided a model that offered a rational explanation for the observed switch in enantioselectivity. (C) 1999 Elsevier Science Ltd. All rights reserved.
|
|
| 5. |
- Eldesokey, Abdelrahman, et al.
(författare)
-
Uncertainty-Aware CNNs for Depth Completion : Uncertainty from Beginning to End
- 2020
-
Ingår i: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). - : IEEE. - 9781728171685 - 9781728171692 ; , s. 12011-12020
-
Konferensbidrag (refereegranskat)abstract
- The focus in deep learning research has been mostly to push the limits of prediction accuracy. However, this was often achieved at the cost of increased complexity, raising concerns about the interpretability and the reliability of deep networks. Recently, an increasing attention has been given to untangling the complexity of deep networks and quantifying their uncertainty for different computer vision tasks. Differently, the task of depth completion has not received enough attention despite the inherent noisy nature of depth sensors. In this work, we thus focus on modeling the uncertainty of depth data in depth completion starting from the sparse noisy input all the way to the final prediction. We propose a novel approach to identify disturbed measurements in the input by learning an input confidence estimator in a self-supervised manner based on the normalized convolutional neural networks (NCNNs). Further, we propose a probabilistic version of NCNNs that produces a statistically meaningful uncertainty measure for the final prediction. When we evaluate our approach on the KITTI dataset for depth completion, we outperform all the existing Bayesian Deep Learning approaches in terms of prediction accuracy, quality of the uncertainty measure, and the computational efficiency. Moreover, our small network with 670k parameters performs on-par with conventional approaches with millions of parameters. These results give strong evidence that separating the network into parallel uncertainty and prediction streams leads to state-of-the-art performance with accurate uncertainty estimates.
|
|
| 6. |
- Gharaee, Zahra, 1986-, et al.
(författare)
-
A Bayesian Approach to Reinforcement Learning of Vision-Based Vehicular Control
- 2021
-
Ingår i: 2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR). - : IEEE COMPUTER SOC. - 9781728188089 ; , s. 3947-3954
-
Konferensbidrag (refereegranskat)abstract
- In this paper, we present a state-of-the-art reinforcement learning method for autonomous driving. Our approach employs temporal difference learning in a Bayesian framework to learn vehicle control signals from sensor data. The agent has access to images from a forward facing camera, which are pre-processed to generate semantic segmentation maps. We trained our system using both ground truth and estimated semantic segmentation input. Based on our observations from a large set of experiments, we conclude that training the system on ground truth input data leads to better performance than training the system on estimated input even if estimated input is used for evaluation. The system is trained and evaluated in a realistic simulated urban environment using the CARLA simulator. The simulator also contains a benchmark that allows for comparing to other systems and methods. The required training time of the system is shown to be lower and the performance on the benchmark superior to competing approaches.
|
|
| 7. |
- Holmberg, Erland, et al.
(författare)
-
Reaction conditions for the resolution of 2-methylalkanoic acids in esterification and hydrolysis with lipase from Candida cylindracea
- 1991
-
Ingår i: Applied Microbiology and Biotechnology. - 0175-7598. ; 35:5, s. 572-578
-
Tidskriftsartikel (refereegranskat)abstract
- We have demonstrated resolution of 2-methylalkanoic acids using lipase from Candida cylindracea as a catalyst. The resolution of 2-methyldecanoic acid was more successful than that of 2-methylbutyric acid both by esterification and hydrolysis. This indicates that the resolution of the acid is dependent on the chain length of the acid moiety. The chain length of the alcohol moiety of the ester affected the resolution of the long-chain acid only. Using esterification, (R)-2-methyldecanoic acid was produced in an enantiomeric excess (e.e.) of 95% (E = 40). If the enantiomeric ratio is low (E = 3.6), as in the resolution of 2-methylbutyric acid, esterification combined with a high equilibrium conversion could be used to yield the remaining acid in a high e.e. In the hydrolytic reactions, the e.e. and the equilibrium conversion were dependent on the pH and the presence of CaCl2. When octyl 2-methyldecanoate was hydrolysed at pH 8.0 in the presence of CaCl2, the (S)-acid was formed with an e.e. of 80% (E = 9), but when the hydrolysis was carried out at pH 7.5 without CaCl2, a very low e.e. and a low equilibrium conversion were observed. The latter conditions allowed the esterification of 2-methyldecanoic acid with 1-octanol even in aqueous medium.
|
|
| 8. |
- Holmquist, Karl, 1992-, et al.
(författare)
-
Class-Incremental Learning for Semantic Segmentation - A study
- 2021
-
Ingår i: 2021 Swedish Artificial Intelligence Society Workshop (SAIS). - : IEEE. - 9781665442367 - 9781665442374 ; , s. 25-28
-
Konferensbidrag (refereegranskat)abstract
- One of the main challenges of applying deep learning for robotics is the difficulty of efficiently adapting to new tasks while still maintaining the same performance on previous tasks. The problem of incrementally learning new tasks commonly struggles with catastrophic forgetting in which the previous knowledge is lost.Class-incremental learning for semantic segmentation, addresses this problem in which we want to learn new semantic classes without having access to labeled data for previously learned classes. This is a problem in industry, where few pre-trained models and open datasets matches exactly the requisites. In these cases it is both expensive and labour intensive to collect an entirely new fully-labeled dataset. Instead, collecting a smaller dataset and only labeling the new classes is much more efficient in terms of data collection.In this paper we present the class-incremental learning problem for semantic segmentation, we discuss related work in terms of the more thoroughly studied classification task and experimentally validate the current state-of-the-art for semantic segmentation. This lays the foundation as we discuss some of the problems that still needs to be investigated and improved upon in order to reach a new state-of-the-art for class-incremental semantic segmentation.
|
|
| 9. |
- Holmquist, Karl, 1992-, et al.
(författare)
-
Computing a Collision-Free Path using the monogenic scale space
- 2018
-
Ingår i: 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). - : IEEE. - 9781538680940 - 9781538680957 - 9781538680933 ; , s. 8097-8102
-
Konferensbidrag (refereegranskat)abstract
- Mobile robots have been used for various purposes with different functionalities which require them to freely move in environments containing both static and dynamic obstacles to accomplish given tasks. One of the most relevant capabilities in terms of navigating a mobile robot in such an environment is to find a safe path to a goal position. This paper shows that there exists an accurate solution to the Laplace equation which allows finding a collision-free path and that it can be efficiently calculated for a rectangular bounded domain such as a map which is represented as an image. This is accomplished by the use of the monogenic scale space resulting in a vector field which describes the attracting and repelling forces from the obstacles and the goal. The method is shown to work in reasonably convex domains and by the use of tessellation of the environment map for non-convex environments.
|
|
| 10. |
- Holmquist, Karl, 1992-
(författare)
-
Data-Driven Robot Perception in the Wild
- 2023
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- As technology continues to advance, the interest in the relief of humans from tedious or dangerous tasks through automation increases. Some of the tasks that have received increasing attention are autonomous driving, disaster relief, and forestry inspection. Developing and deploying an autonomous robotic system to this type of unconstrained environments —in a safe way— is highly challenging. The system requires precise control and high-level decision making. Both of which require a robust and reliable perception system to understand the surroundings correctly. The main purpose of perception is to extract meaningful information from the environment, be it in the form of 3D maps, dense classification of the type of object and surfaces, or high-level information about the position and direction of moving objects. Depending on the limitations and application of the system, various types of sensors can be used: lidars, to collect sparse depth information; cameras, to collect dense information for different parts of the visual spectra, of-ten the red-green-blue (RGB) bands; Inertial Measurements Units (IMUs), to estimate the ego motion; microphones, to interact and respond to humans; GPS receivers, to get global position information; just to mention a few. This thesis investigates some of the necessities to approach the requirements of this type of system. Specifically, focusing on data-driven approaches, that is, machine learning, which has been shown time and again to be the main competitor for high-performance perception tasks in recent years. Although precision requirements might be high in industrial production plants, the environment is relatively controlled and the task is fixed. Instead, this thesis is studying some of the aspects necessary for complex, unconstrained environments, primarily outdoors and potentially near humans or other systems. The term in the wild refers exactly to the unconstrained nature of these environments, where the system can easily encounter something previously unseen and where the system might interact with unknowing humans. Some examples of environments are: city traffic, disaster relief scenarios, and dense forests. This thesis will mainly focus on the following three key aspects necessary to handle the types of tasks and situations that could occur in the wild: 1) generalizing to a new environment, 2) adapting to new tasks and requirements, and 3) modeling uncertainty in the perception system. First, a robotic system should be able to generalize to new environments and still function reliably. Papers B and G address this by using an intermediate representation to allow the system to handle much more diverse types of environment than otherwise possible. Paper B also investigates how robust the proposed autonomous driving system was to incorrect predictions, which is one of the likely results of changing the environment. Second, a robot should be sufficiently adaptive to allow it to learn new tasks without forgetting the previous ones. Paper E proposed a way to allow incrementally adding new semantic classes to a trained model without access to the previous training data. The approach is based on utilizing the uncertainty in the predictions to model the unknown classes, marked as background. Finally, the perception system will always be partially flawed, either because of the lack of modeling capabilities or because of ambiguities in the sensor data. To properly take this into account, it is fundamental that the system has the ability to estimate the certainty in the predictions. Paper F proposed a method for predicting the uncertainty in the model predictions when interpolating sparse data. Paper G addresses the ambiguities that exist when estimating the 3D pose of a human from a single camera image.
|
|
| 11. |
- Holmquist, Karl, et al.
(författare)
-
DiffPose: Multi-hypothesis Human Pose Estimation using Diffusion Models
- 2023
-
Ingår i: 2023 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2023). - : IEEE COMPUTER SOC. - 9798350307184 - 9798350307191 ; , s. 15931-15941
-
Konferensbidrag (refereegranskat)abstract
- Traditionally, monocular 3D human pose estimation employs a machine learning model to predict the most likely 3D pose for a given input image. However, a single image can be highly ambiguous and induces multiple plausible solutions for the 2D-3D lifting step, which results in overly confident 3D pose predictors. To this end, we propose DiffPose, a conditional diffusion model that predicts multiple hypotheses for a given input image. Compared to similar approaches, our diffusion model is straightforward and avoids intensive hyperparameter tuning, complex network structures, mode collapse, and unstable training. Moreover, we tackle the problem of over-simplification of the intermediate representation of the common two-step approaches which first estimate a distribution of 2D joint locations via joint-wise heatmaps and consecutively use their maximum argument for the 3D pose estimation step. Since such a simplification of the heatmaps removes valid information about possibly correct, though labeled unlikely, joint locations, we propose to represent the heatmaps as a set of 2D joint candidate samples. To extract information about the original distribution from these samples, we introduce our embedding transformer which conditions the diffusion model. Experimentally, we show that DiffPose improves upon the state of the art for multi-hypothesis pose estimation by 3-5% for simple poses and outperforms it by a large margin for highly ambiguous poses.(1)
|
|
| 12. |
- Holmquist, Karl, 1992-, et al.
(författare)
-
Evidential Deep Learning for Class-Incremental Semantic Segmentation
- 2023
-
Ingår i: Image Analysis. SCIA 2023.. - : Springer. - 9783031314377 - 9783031314384 ; , s. 32-48
-
Konferensbidrag (refereegranskat)abstract
- Class-Incremental Learning is a challenging problem in machine learning that aims to extend previously trained neural networks with new classes. This is especially useful if the system is able to classify new objects despite the original training data being unavailable. Although the semantic segmentation problem has received less attention than classification, it poses distinct problems and challenges, since previous and future target classes can be unlabeled in the images of a single increment. In this case, the background, past and future classes are correlated and there exists a background-shift.In this paper, we address the problem of how to model unlabeled classes while avoiding spurious feature clustering of future uncorrelated classes. We propose to use Evidential Deep Learning to model the evidence of the classes as a Dirichlet distribution. Our method factorizes the problem into a separate foreground class probability, calculated by the expected value of the Dirichlet distribution, and an unknown class (background) probability corresponding to the uncertainty of the estimate. In our novel formulation, the background probability is implicitly modeled, avoiding the feature space clustering that comes from forcing the model to output a high background score for pixels that are not labeled as objects. Experiments on the incremental Pascal VOC and ADE20k benchmarks show that our method is superior to the state of the art, especially when repeatedly learning new classes with increasing number of increments.
|
|
| 13. |
- Holmquist, Mats Torsten, et al.
(författare)
-
Lipases from Rhizomucor miehei and Humicola lanuginosa : Modification of the lid covering the active site alters enantioselectivity
- 1993
-
Ingår i: Journal of Protein Chemistry. - 0277-8033. ; 12:6, s. 749-757
-
Tidskriftsartikel (refereegranskat)abstract
- The homologous lipases from Rhizomucor miehei and Humicola lanuginosa showed approximately the same enantioselectivity when 2-methyldecanoic acid esters were used as substrates. Both lipases preferentially hydrolyzed the S- enantiomer of 1-heptyl 2-methyldecanoate (R. miehei: E(S) = 8.5; H. lanuginosa: E(S) = 10.5), but the R-enantiomer of phenyl 2-methyldecanoate (E(R) = 2.9). Chemical arginine specific modification of the R. miehei lipase with 1,2-cyclohexanedione resulted in a decreased enantioselectivity (E(R) = 2.0), only when the phenyl ester was used as a substrate. In contrast, treatment with phenylglyoxal showed a decreased enantioselectivity (E(S) = 2.5) only when the heptyl ester was used as a substrate. The presence of guanidine, an arginine side chain analog, decreased the enantioselectivity with the heptyl ester (E(S) = 1.9) and increased the enantioselectivity with the aromatic ester (E(R) = 4.4) as substrates. The mutation, Glu 87 Ala, in the lid of the H. lanuginosa lipase, which might decrease the electrostatic stabilization of the open-lid conformation of the lipase, resulted in 47% activity compared to the native lipase, in a tributyrin assay. The Glu 87 Ala mutant showed an increased enantioselectivity with the heptyl ester (E(S) = 17.4) and a decreased enantioselectivity with the phenyl ester (E(R) = 2.5) as substrates, compared to native lipase. The enantioselectivities of both lipases in the esterification of 2-methyldecanoic acid with 1-heptanol were unaffected by the lid modifications.
|
|
| 14. |
- Klamt, Tobias, et al.
(författare)
-
Flexible Disaster Response of Tomorrow: Final Presentation and Evaluation of the CENTAURO System
- 2019
-
Ingår i: IEEE robotics & automation magazine. - : IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC. - 1070-9932 .- 1558-223X. ; 26:4, s. 59-72
-
Tidskriftsartikel (refereegranskat)abstract
- Mobile manipulation robots have great potential for roles in support of rescuers on disaster-response missions. Robots can operate in places too dangerous for humans and therefore can assist in accomplishing hazardous tasks while their human operators work at a safe distance. We developed a disaster-response system that consists of the highly flexible Centauro robot and suitable control interfaces, including an immersive telepresence suit and support-operator controls offering different levels of autonomy.
|
|
| 15. |
- Lilja, Karl, et al.
(författare)
-
Bedömning av miljögiftspåverkan i vattenmiljö. Samordnad metodutveckling
- 2010
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- Ett förslag till vägledning för att bedriva miljögiftsövervakning och för hur framtagna data skall kunna användas i vattenförvaltningen har inom denna studie tagits fram. Förslaget till vägledning syftar också till att ge underlag för en sammanhållen metodik för hur existerande datakällor på ett effektivt sätt skall kunna utnyttjas vid bedömning av miljögiftspåverkan i vattendirektivsarbetet. Förslag och rekommendationer har baserats på krav som ställs enligt EUs vattendirektiv och har kompletteras utifrån resultaten i föreliggande studie. En kritisk granskning av de översättningar mellan EQS för ytvatten, och motsvarande halter i sediment och biota, som tidigare har utförts med hjälp av jämviktsfördelningsmetodik och bioackumuleringsfaktorer enligt den metodik som beskrivs i TGD (Technical Guidance Document on Risk Assessment) (TGD, 2003) har utförts för att bedöma huruvida metoden är tillämplig i svenska förhållanden. Halter av miljögifter i sediment, biota och vatten har jämförts med gränsvärden, antingen beräknade med översättningsmetodik eller uppsatta i direktiv
|
|
| 16. |
- Lilja, Karl, et al.
(författare)
-
Chemical and biological monitoring of sewage effluent water
- 2010
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- As an assignment from the Swedish Environmental Protection Agency, IVL has conducted a study with the overall aim to present basic data for decision-making regarding a monitoring program for STP effluents. A program for chemical and biological monitoring in effluent water from municipal sewage treatment plants is important for the understanding on spreading of chemicals in the environment, to identify changes in the use or substitution of chemicals, to identify possible emissions of 'new' emerging substances to the environment, and also for the implementation and follow up of directives. The study consisted of several parts; a chemical characterization of STP effluents consisting of both the analysis of specific compounds but also the identification of 'unknown' compounds, measurements of estrogenic and androgenic activity with the bioassays YES and YAS, evaluation of seasonal variability in chemical composition, and chemical measurements in recipient water. Further, a literature survey of potential additional bioassays suitable for monitoring purposes was conducted. Based on the results a recommendation for a monitoring program for municipal STP effluent is given.
|
|
| 17. |
|
|
| 18. |
- Park, S., et al.
(författare)
-
Focusing mutations into the P. fluorescens esterase binding site increases enantioselectivity more effectively than distant mutations
- 2005
-
Ingår i: Chemistry and Biology. - : Elsevier BV. - 1074-5521 .- 1879-1301. ; 12:1, s. 45-54
-
Tidskriftsartikel (refereegranskat)abstract
- Rational design of enzymes with improved properties, such as enantioselectivity, usually focuses mutations within the substrate binding site. On the other hand, directed evolution of enzymes usually targets the entire protein and discovers beneficial mutations far from the substrate binding site. In this paper, we propose an explanation for this discrepancy and show that a combined approach-random mutagenesis within the substrate binding site-is better. To increase the enantioselectivity (E) of a Pseudomonas fluorescens esterase (PFE) toward methyl 3-bromo-2-methylpropionate, we focused mutagenesis into the substrate binding site at Trp28, Val121, Phe198, and Val225. Five of the catalytically active mutants (13%) showed better enantioselectivity than wild-type PFE. The increases in enantioselectivity were higher (up to 5-fold, reaching E = 61) than with mutants identified by random mutagenesis of the entire enzyme.
|
|
| 19. |
- Rotticci-Mulder, J. C., et al.
(författare)
-
Expression in Pichia pastoris of Candida antarctica lipase B and lipase B fused to a cellulose-binding domain
- 2001
-
Ingår i: Protein Expression and Purification. - : Elsevier BV. - 1046-5928 .- 1096-0279. ; 21:3, s. 386-392
-
Tidskriftsartikel (refereegranskat)abstract
- Candida antarctica lipase B (CALB) and C. antarctica lipase B fused to a cellulose-binding domain (CBD-CALB) were expressed functionally in the methylotrophic yeast Pichia pastoris. The cellulose-binding domain originates from cellulase A of the anaerobic rumen fungus Neocallimastix patriciarum. The genes were fused to the a-factor secretion signal sequence of Saccharomyces cerevisiae and placed under the control of the alcohol oxidase gene (AOX1) promoter. The recombinant proteins were secreted into the culture medium reaching levels of approximately 25 mg/L. The proteins were purified using hydrophobic interaction chromatography and gel filtration with an overall yield of 69%. Results from endoglycosidase H digestion of the proteins showed that CALB and CBD-CALB were N-glycosylated. The specific hydrolytic activities of recombinant CALB and CBD-CALB were identical to that reported for CALB isolated from its native source. The fusion of the CBD to the lipase resulted in a greatly enhanced binding toward cellulose for CBD-CALB compared with that for CALB.
|
|
| 20. |
|
|
| 21. |
- Wennberg, Karl, et al.
(författare)
-
Problemistic search and international entrepreneurship
- 2008
-
Ingår i: European Management Journal. - : Elsevier. - 0263-2373. ; 26:6, s. 441-454
-
Tidskriftsartikel (refereegranskat)abstract
- This paper explains the internationalization process of small firms using the theory of performance relative to aspiration levels. The study complements prior theory by explaining why and how small firms are triggered to engage in internationalization despite not reaching maturity in their home market. We outline a model where firms’ internationalization is triggered by problemistic search, following periods of below-aspiration performance. The model is tested on 860 Swedish firms followed during an economic downturn. Results indicate that internationalization activities follow a bounded rational process characterized by search behavior which is triggered by performance feedback. The study complements prior theories of internationalization and offers a first empirical demonstration of the viability of aspiration-level performance theory in international entrepreneurship research.
|
|
