ACE research vignette 016 : balanced skills - effects and origins

This series of research vignettes is aimed at sharing current and interesting research findings from our team of international Entrepreneurship researchers. In this vignette, post-doctoral research fellow Dr Michael Stuetzer considers the effects and origins of balanced skills.

Do balanced skills help nascent entrepreneurs to make progress in the venture creation process?

We analyze longitudinal data on innovative start-up projects and apply Lazear’s jack-of-all-trades theory to investigate the effect of nascent entrepreneurs’ balanced skills on their progress in the venture creation process. Our results suggest that those nascent entrepreneurs who exhibit a sufficiently broad set of skills undertake more gestation activities towards an operational new venture. This supports the notion that a balanced skill set is an important determinant of entrepreneurial market entry.

Balanced skills among nascent entrepreneurs

This paper examines the effects and origins of balanced skills among nascent entrepreneurs. In a first step we apply Lazear’s jack-of-all-trades theory to investigate performance effects of a balanced skill set. Second, we investigate potential sources of balanced skills, thereby testing the investment hypothesis against the endowment hypothesis. Analyzing data on high-potential nascent projects, we find support for the notion that balanced skills are important for making progress in the venture creation process. Regarding the origins of balanced skills, the data support both hypotheses. In line with the investment hypothesis an early interest in an entrepreneurial career, prior managerial and entrepreneurial experience are significantly related with a more balanced skill set. Supporting the endowment hypothesis, an entrepreneurial personality profile indicating entrepreneurial talent is correlated with a balanced skill set. Our results thus hint at the need for theories on the origins of a balanced skill set that integrate both views.

A tensor encoding model for semantic processing

This paper develops and evaluates an enhanced corpus based approach for semantic processing. Corpus based models that build representations of words directly from text do not require pre-existing linguistic knowledge, and have demonstrated psychologically relevant performance on a number of cognitive tasks. However, they have been criticised in the past for not incorporating sufficient structural information. Using ideas underpinning recent attempts to overcome this weakness, we develop an enhanced tensor encoding model to build representations of word meaning for semantic processing. Our enhanced model demonstrates superior performance when compared to a robust baseline model on a number of semantic processing tasks.

Semantic judgement of medical concepts : combining syntagmatic and paradigmatic information with the tensor encoding model

This paper outlines a novel approach for modelling semantic relationships within medical documents. Medical terminologies contain a rich source of semantic information critical to a number of techniques in medical informatics, including medical information retrieval. Recent research suggests that corpus-driven approaches are effective at automatically capturing semantic similarities between medical concepts, thus making them an attractive option for accessing semantic information. Most previous corpus-driven methods only considered syntagmatic associations. In this paper, we adapt a recent approach that explicitly models both syntagmatic and paradigmatic associations. We show that the implicit similarity between certain medical concepts can only be modelled using paradigmatic associations. In addition, the inclusion of both types of associations overcomes the sensitivity to the training corpus experienced by previous approaches, making our method both more effective and more robust. This finding may have implications for researchers in the area of medical information retrieval.

On being balanced in an unbalanced world

This paper examines the case of a procurement auction for a single project, in which the breakdown of the winning bid into its component items determines the value of payments subsequently made to bidder as the work progresses. Unbalanced bidding, or bid skewing, involves the uneven distribution of mark-up among the component items in such a way as to attempt to derive increased benefit to the unbalancer but without involving any change in the total bid. One form of unbalanced bidding for example, termed Front Loading (FL), is thought to be widespread in practice. This involves overpricing the work items that occur early in the project and underpricing the work items that occur later in the project in order to enhance the bidder's cash flow. Naturally, auctioners attempt to protect themselves from the effects of unbalancing—typically reserving the right to reject a bid that has been detected as unbalanced. As a result, models have been developed to both unbalance bids and detect unbalanced bids but virtually nothing is known of their use, success or otherwise. This is of particular concern for the detection methods as, without testing, there is no way of knowing the extent to which unbalanced bids are remaining undetected or balanced bids are being falsely detected as unbalanced. This paper reports on a simulation study aimed at demonstrating the likely effects of unbalanced bid detection models in a deterministic environment involving FL unbalancing in a Texas DOT detection setting, in which bids are deemed to be unbalanced if an item exceeds a maximum (or fails to reach a minimum) ‘cut-off’ value determined by the Texas method. A proportion of bids are automatically and maximally unbalanced over a long series of simulated contract projects and the profits and detection rates of both the balancers and unbalancers are compared. The results show that, as expected, the balanced bids are often incorrectly detected as unbalanced, with the rate of (mis)detection increasing with the proportion of FL bidders in the auction. It is also shown that, while the profit for balanced bidders remains the same irrespective of the number of FL bidders involved, the FL bidder's profit increases with the greater proportion of FL bidders present in the auction. Sensitivity tests show the results to be generally robust, with (mis)detection rates increasing further when there are fewer bidders in the auction and when more data are averaged to determine the baseline value, but being smaller or larger with increased cut-off values and increased cost and estimate variability depending on the number of FL bidders involved. The FL bidder's expected benefit from unbalancing, on the other hand, increases, when there are fewer bidders in the auction. It also increases when the cut-off rate and discount rate is increased, when there is less variability in the costs and their estimates, and when less data are used in setting the baseline values.

On robust face recognition via sparse encoding : the good, the bad, and the ugly

In the field of face recognition, Sparse Representation (SR) has received considerable attention during the past few years. Most of the relevant literature focuses on holistic descriptors in closed-set identification applications. The underlying assumption in SR-based methods is that each class in the gallery has sufficient samples and the query lies on the subspace spanned by the gallery of the same class. Unfortunately, such assumption is easily violated in the more challenging face verification scenario, where an algorithm is required to determine if two faces (where one or both have not been seen before) belong to the same person. In this paper, we first discuss why previous attempts with SR might not be applicable to verification problems. We then propose an alternative approach to face verification via SR. Specifically, we propose to use explicit SR encoding on local image patches rather than the entire face. The obtained sparse signals are pooled via averaging to form multiple region descriptors, which are then concatenated to form an overall face descriptor. Due to the deliberate loss spatial relations within each region (caused by averaging), the resulting descriptor is robust to misalignment & various image deformations. Within the proposed framework, we evaluate several SR encoding techniques: l1-minimisation, Sparse Autoencoder Neural Network (SANN), and an implicit probabilistic technique based on Gaussian Mixture Models. Thorough experiments on AR, FERET, exYaleB, BANCA and ChokePoint datasets show that the proposed local SR approach obtains considerably better and more robust performance than several previous state-of-the-art holistic SR methods, in both verification and closed-set identification problems. The experiments also show that l1-minimisation based encoding has a considerably higher computational than the other techniques, but leads to higher recognition rates.

A tensor encoding model of word meaning : theory and application to information retrieval

This project was a step forward in developing and evaluating a novel, mathematical model that can deduce the meaning of words based on their use in language. This model can be applied to a wide range of natural language applications, including the information seeking process most of us undertake on a daily basis.

Some extensions to representation and encoding of structure in models of distributional semantics

This thesis makes several contributions towards improved methods for encoding structure in computational models of word meaning. New methods are proposed and evaluated which address the requirement of being able to easily encode linguistic structural features within a computational representation while retaining the ability to scale to large volumes of textual data. Various methods are implemented and evaluated on a range of evaluation tasks to demonstrate the effectiveness of the proposed methods.

“Freer than we Feel” : Applying the Critical Dimension of Consciousness to Children’s Multimodal Encoding

Reports of children and teachers taking transformative social action in schools are becoming rare. This session illustrates how teachers, while feeling the weight of accountability testing in schools, are active agents who can re-imagine literacy pedagogy to change elements of their community. It reports the critical dimensions of a movie-making unit with Year 5 students within a school reform project. The students filmed interviews with people in the local shops to gather lay-knowledge and experiences of the community. The short documentaries challenged stereotypes about what it is like to live in Logan, and critically identified potential improvements to public spaces in the local community. A student panel presented these multimodal texts at a national conference of social activists and community leaders. The report does not valorize or privilege local or lay knowledge over dominant knowledge, but argues that prescribed curriculum should not hinder the capacity for critical consciousness.

Small RNA sequencing of potato leafroll virus-infected plants reveals an additional subgenomic RNA encoding a sequence-specific RNA-binding protein

Potato leafroll virus (PLRV) is a positive-strand RNA virus that generates subgenomic RNAs (sgRNA) for expression of 3' proximal genes. Small RNA (sRNA) sequencing and mapping of the PLRV-derived sRNAs revealed coverage of the entire viral genome with the exception of four distinctive gaps. Remarkably, these gaps mapped to areas of PLRV genome with extensive secondary structures, such as the internal ribosome entry site and 5' transcriptional start site of sgRNA1 and sgRNA2. The last gap mapped to ~500. nt from the 3' terminus of PLRV genome and suggested the possible presence of an additional sgRNA for PLRV. Quantitative real-time PCR and northern blot analysis confirmed the expression of sgRNA3 and subsequent analyses placed its 5' transcriptional start site at position 5347 of PLRV genome. A regulatory role is proposed for the PLRV sgRNA3 as it encodes for an RNA-binding protein with specificity to the 5' of PLRV genomic RNA. © 2013.

A single copy of a virus-derived transgene encoding hairpin RNA gives immunity to barley yellow dwarf virus

Barley yellow dwarf virus-PAV (BYDV-PAV) is the most serious and widespread virus of cereals worldwide. Natural resistance genes against this luteovirus give inadequate control, and previous attempts to introduce synthetic resistance into cereals have produced variable results. In an attempt to generate barley with protection against BYDV-PAV, plants were transformed with a transgene designed to produce hairpin (hp)RNA containing BYDV-PAV sequences. From 25 independent barley lines transformed with the BYDV-PAV hpRNA construct, nine lines showed extreme resistance to the virus and the majority of these contained a single transgene. In the progeny of two independent transgenic lines, inheritance of a single transgene consistently correlated with protection against BYDV-PAV. This protection was rated as immunity because the virus could not be detected in the challenged plants by ELISA nor recovered by aphid feeding experiments. In the field, BYDV-PAV is sometimes associated with the related luteovirus Cereal yellow dwarf virus-RPV (CYDV-RPV). When the transgenic plants were challenged with BYDV-PAV and CYDV-RPV together, the plants were susceptible to CYDV-RPV but immune to BYDV-PAV. This shows that the immunity is virus-specific and not broken down by the presence of CYDV. It suggests that CYDV-RPV does not encode a silencing-suppressor gene or that its product does not protect BYDV-PAV against the plant's RNAi-like defence mechanism. Either way, our results indicate that the BYDV-PAV immunity will be robust in the field and is potentially useful in minimizing losses in cereal production worldwide.

Developing a framework for subcontractor appraisal using a balanced scorecard

Subcontractors contribute significantly to construction projects and their performance can seriously affect over-all project success. It is crucial, therefore, to appraise the performance of subcontractors to ensure they satisfy the client's expectations and project requirements. To increase the transparency and accuracy of subcontractor appraisal, the baseline and target performance levels should be set at the outset so that the appraisers and those being appraised realize exactly what standards are to be achieved. The balanced scorecard, being a powerful tool for performance appraisal, offers a potentially good approach for modeling the subcontractor appraisal process. In this paper, an approach to developing a balanced scorecard subcontractor appraisal model is proposed and demonstrated through a questionnaire survey administered in Hong Kong and from which the baseline and target performance levels for large-scale skilled subcontractors are identified. A case example is used to illustrate the operation of the model. Finally, a means by which the model may be validated is demonstrated through the use of field experts. The results demonstrate the feasibility of developing a balanced scorecard model that can help improve the transparency of subcontractor appraisal despite the baseline and target performance levels changing according to the project characteristics, subcontractor categories and size.

A gene encoding a new cold-active lipase from an Antarctic isolate of Penicillium expansum

Cold-active lipases are of significant interest as biocatalysts in industrial processes. We have identified a lipase that displayed activity towards long carbon-chain-p-nitrophenyl substrates (C12–C18) at 25 °C from the culture supernatant of an Antarctic Penicillium expansum strain assigned P. expansum SM3. Zymography revealed a protein band of around 30 kDa with activity towards olive oil. DNA fragments of a lipase gene designated as lipPE were isolated from the genomic DNA of P. expansum SM3 by genomic walking PCR. Subsequently, the complete genomic lipPE gene was amplified using gene-specific primers designed from the 5′- and 3′-regions. Reverse transcription PCR was used to amplify the lipPE cDNA. The deduced amino acid sequence consisted of 285 residues that included a predicted signal peptide. Three peptides identified by LC/MS/MS analysis of the proteins in the culture supernatant of P. expansum were also present in the deduced amino acid sequence of the lipPE gene suggesting that this gene encoded the lipase identified by initial zymogram activity analysis. Full analysis of the nucleotide and the deduced amino acid sequences indicated that the lipPE gene encodes a novel P. expansum lipase. The lipPE gene was expressed in E. coli for further characterization of the enzyme with a view of assessing its suitability for industrial applications.

The organisation and expression of the genes encoding the mitochondrial glycine decarboxylase complex and serine hydroxymethyltransferase in pea (Pisum sativum)

Restriction fragment length polymorphisms have been used to determine the chromosomal location of the genes encoding the glycine decarboxylase complex (GDC) and serine hydroxymethyltransferase (SHMT) of pea leaf mitochondria. The genes encoding the H subunit of GDC and the genes encoding SHMT both show linkage to the classical group I marker i. In addition, the genes for the P protein of GDC show linkage to the classic group I marker a. The genes for the L and T proteins of GDC are linked to one another and are probably situated on the satellite of chromosome 7. The mRNAs encoding the five polypeptides that make up GDC and SHMT are strongly induced when dark-grown etiolated pea seedlings are placed in the light. Similarly, when mature plants are placed in the dark for 48 h, the levels of both GDC protein and SHMT mRNAs decline dramatically and then are induced strongly when these plants are returned to the light. During both treatments a similar pattern of mRNA induction is observed, with the mRNA encoding the P protein of GDC being the most rapidly induced and the mRNA for the H protein the slowest. Whereas during the greening of etiolated seedlings the polypeptides of GDC and SHMT show patterns of accumulation similar to those of the corresponding mRNAs, very little change in the level of the polypeptides is seen when mature plants are placed in the dark and then re-exposed to the light.