Critical Choices Research Articles

Making Human Learning Visible in a World of Invisible Generative AI

The growth of Generative Artificial Intelligence (GenAI) in Higher Education (HE) challenges us to think differently and creatively about learning and assessment in order to prepare graduates for their futures (Dawson, 2023) This presentation will argue that the very strength of GenAI is also its Achilles’ heel. Anyone can access, garner and generate data rapidly, but human beings have never had a greater need for scrutiny and evaluation of this output. Our graduates will need to have a transparent understanding of how they are learning in terms of the critical choices they make, the information or knowledge they engage with, and where machine learning ends and theirs begins (Okolie et al, 2022). The future of assessment in HE must go beyond simply measuring knowledge acquisition and instead focus on capturing and evidencing students’ capacity for critical thinking. We will highlight some preliminary findings and early impacts of work underway at the University of Bath (UK) and University of Stellenbosch (South Africa), who are working collaboratively to ensure students develop holistically, excel in critical thinking, and gain the confidence and skills to use GenAI effectively to advance their ideas and learning (Osborne, 2024). This will be structured in three parts: Part one (Reviewing): We will outline work during academic year 23/24 by academic staff to review all their assessments against a three-fold categorisation model of permitted GenAI usage. From September 2024, all assessments clearly signal to students the extent to which GenAI is permitted in each assignment. Part two (Guidance): We will showcase resources that support staff to rethink assessment design that embeds GenAI in ways that are authentic, valid and robust (Upsher et al, 2024). We will also present resources that support students to consider how they might use GenAI ethically, and to develop their AI literacy skills and critical thinking. Part three (Examples): We will present initial findings from an internal project that is engaging with national and international employers, in-house world-leading AI researchers and our own students that highlight creative approaches to harnessing GenAI as a tool to reinforce and support (rather than replace) human learning. We conclude by suggesting that HE has always promoted critical thinking and the rigorous investigation and examination of the production of knowledge. We need graduates who can critically engage with their learning, and assessments that capture and evidence this. The only way to do this effectively is to put people at the start, middle and end of our assessment journey.

Seasonal forecasts have sufficient skill to inform some agricultural decisions.

Abstract Seasonal forecasts, which look several months into the future, are currently underutilized in active decision-making, particularly for agricultural and natural resource management. This underutilization can be attributed to the absence of forecasts for decision-relevant variables at the required spatiotemporal resolution and at the time when the decisions are made and a perception of poor skill by decision-makers. Addressing these constraints, we quantified the skill of seasonal forecasts in informing two agricultural decisions with differing decision timeframes and influencer variables: (a) whether to apply fertilizer in fall or wait until spring based on expected winter temperatures, and (b) drought response, such as whether to lease water based on expectations of drought. We also looked into how early the forecast can be provided without significant degradation in skill. Currently, drought response decisions are typically formulated in April, utilizing drought forecasts issued in the same month, while fall fertilization decisions are generally made between August and September. There is growing interest among stakeholders in the availability of earlier forecasts to inform these critical choices. We utilized the North American Multi-Model Ensemble (NMME) hindcasts for the time period 1982 - 2020 over the Pacific Northwest US to obtain meteorological variables. Runoff was estimated via simulations of the coupled crop-hydrology VIC-CropSyst model. The skill assessment with the Heidke Skill Score (HSS) yielded promising outcomes in both decisions for the entire Pacific Northwest region. Notably, NMME’s positive skill (median HSS of 30%) in predicting warmer winters identifies years when fertilizer application should be avoided to prevent fertilizer loss through mineralization (and associated costs). Similarly, there is skill in forecasting drought conditions in most irrigated watersheds for up to two months in advance of April, the current decision time. In conclusion, our findings affirm that contrary to the perception of low skill and resulting underutilization, current seasonal forecasts hold the potential to inform at least some key agricultural decisions.

Transient chromatin decompaction at the start of D. melanogaster male embryonic germline development.

Embryonic germ cells develop rapidly to establish the foundation for future developmental trajectories, and in this process, they make critical lineage choices including the configuration of their unique identity and a decision on sex. Here, we use single-cell genomics patterns for the entire embryonic germline in Drosophila melanogaster along with the somatic gonadal precursors after embryonic gonad coalescence to investigate molecular mechanisms involved in the setting up and regulation of the germline program. Profiling of the early germline chromatin landscape revealed sex- and stage-specific features. In the male germline immediately after zygotic activation, the chromatin structure underwent a brief remodeling phase during which nucleosome density was lower and deconcentrated from promoter regions. These findings echoed enrichment analysis results of our genomics data in which top candidates were factors with the ability to mediate large-scale chromatin reorganization. Together, they point to the importance of chromatin regulation in the early germline and raise the possibility of a conserved epigenetic reprogramming-like process required for proper initiation of germline development.

Net Present Value Calculations for Mining Post-Closure Financial Assurance

AbstractMost regulatory jurisdictions now require financial assurance for mines for post-closure activities, which typically include long-term monitoring and maintenance, and most importantly post-closure water treatment. Critical choices in calculating post-closure financial assurance include using a reputable cost estimation model to establish the amount needed for reclamation, selecting a realistic net return-on-investment for post-closure trust funds, and choosing the period of time over which the net present value calculation of the financial assurance is based. A conservative estimate of the amount of financial assurance is warranted to protect the public from assuming these costs. However, some regulatory agencies use financial assurance estimate practices that are arbitrary. The data and procedures needed to analyze and recommend conservative net present value calculations are readily available, but in too many instances are not being utilized.

CEO power and risk taking: evidence from European banks mergers and acquisitions

PurposeThis study investigates the relationship between CEO power and the risk profile (RP) of acquiring banks through mergers and acquisitions (M&A) transactions.Design/methodology/approachThe analysis is based on 214 transactions between 2010 and 2022 involving European Union-based acquirers. To assess the impact of M&A on the acquiring bank’s RP, we compare changes in the acquirer’s RP to control banks. We use linear regression with two-stage least squares instrumental variables (2SLS-IV) to examine the effect of CEO power on changes in merger-related risk.FindingsThe findings suggest that CEO power reduces the RP of the acquiring bank. Specifically, CEOs who hold both the CEO and board chair positions tend to take fewer risks. Additionally, CEOs with high ownership, CEO pay and extensive experience (measured by tenure and acquisition experience) decrease the RP. However, prestige power is positively correlated with an increase in RPs.Practical implicationsThis research examines CEO influence on bank risk post-mergers, providing insights into governance, risk and strategic choices. The findings can guide banks in CEO selection and governance to mitigate M&A risks, improving risk management and decision-making in the financial sector.Originality/valueThis study is the first empirical investigation introducing diverse executive power metrics to analyze the link between executive power and risk-taking in the European banking sector, with a specific emphasis on the impact on M&A as critical investment choices.

Introducing comprehensive sexuality education in Italian schools: from the co-construction to the evaluation of a pilot intervention

IntroductionSchool-based comprehensive sexuality education (CSE) is a powerful tool that provides young people with information on all aspects of sexuality and is aimed at protecting their sexual and reproductive health and well-being throughout their lives. Currently, CSE is not integrated within the schools’ curriculum in Italy. This study describes the co-construction, implementation, and evaluation of a CSE project piloted among students attending lower secondary schools, in four regions of Italy. Evidence-based evaluation will be helpful in promote the inclusion of CSE programs in the Italian schools’ curriculum.MethodsThe pilot scheme was co-constructed by a multidisciplinary curriculum development group through a Delphi process, including educators who conducted the activities. The evaluation followed three directions: the program (based on a literature review of CSE principles and recommended characteristics), implementation (assessing the execution of the program through the analysis of the reflection tools used by the educators), and short-term outcomes (assessing critical thinking and conscious behavioural choices through pre-post and satisfaction surveys).ResultsThe main goal, learning modules and content were defined and structured in five interventions with the students, and two with families and teachers. A total of 638 students were involved in the activity, across 11 schools. Data analysis of pre/post surveys reported a significant increase in knowledge in 12 of the 15 items investigated (p < 0.05), and a high level of satisfaction with the topics addressed. Qualitative analysis added information on the pivotal role of educators in CSE.ConclusionsThe national piloting of this educational activity provided positive insights regarding the co-construction, implementation and short-outcome evaluation, suggesting potential for scalability and future inclusion of CSE in the curricula of Italian schools.

Normative Challenges of Risk Regulation of Artificial Intelligence

Approaches aimed at regulating artificial intelligence (AI) include a particular form of risk regulation, i.e. a risk-based approach. The most prominent example is the European Union’s Artificial Intelligence Act (AI Act). This article addresses the challenges for adequate risk regulation that arise primarily from the specific type of risks involved, i.e. risks to the protection of fundamental rights and fundamental societal values. This is mainly due to the normative ambiguity of such rights and societal values when attempts are made to select, interpret, specify or operationalise them for the purposes of risk assessments and risk mitigation. This is exemplified by (1) human dignity, (2) informational self-determination, data protection and privacy, (3) anti-discrimination, fairness and justice, and (4) the common good. Normative ambiguities require normative choices, which are assigned to different actors under the regime of the AI Act. Particularly critical normative choices include selecting normative concepts by which to operationalise and specify risks, aggregating and quantifying risks (including the use of metrics), balancing value conflicts, setting levels of acceptable risks, and standardisation. To ensure that these normative choices do not lack democratic legitimacy and to avoid legal uncertainty, further political processes and scientific debates are suggested.

Investigating off-road vehicle lateral stability with integrated chassis control

This study investigates the improvement of off-road vehicle lateral stability by integrated control of active rear steering (ARS) and rear differential braking (RDB) and how the performance of such systems compares on smooth and rough roads. The ARS and RDB controllers comprise a sliding mode controller (SMC) for which critical design choices are the SMC reference model, SMC gain and integration rule. Findings include that the kinematic model reference error is preferred over the phase plane location error on both terrains, the SMC gain is terrain dependent, and the rear axle slip angle is the preferred integration rule over the stability index (SI) on both terrains. The study also found that RDB, and to a lesser degree ARS, tends to improve on the baseline vehicle path-following ability for a double lane change (DLC) manoeuvre on both terrains, but RDB has a larger loss of speed compared to ARS. The Rear axle slip angle was found to be a terrain-dependent tuneable integration rule to combine ARS and RDB, and this resulted in a control system that has the good path-following ability of RDB but the low loss of speed associated with ARS after tuning.

Australia’s national(ist) history curriculum: history education as a site of attempted de-democratisation

AbstractThis paper explores the contested content of the Australian history curriculum to understand the curriculum’s national(ist?) purpose and investigate if national histories can be taught in a way which combats the anti-democratic forces at play in our culture. This question will be explored through analysis of the three topics in the Australian Curriculum: History 7–10, which have a strong focus on Australian history specifically, and semi-structured interviews with secondary teachers on pedagogies for history and instilling democratic dispositions in students (UniSQ ETH2023-0315). Since Prime Minister John Howard’s call for reform of the curriculum to ensure that the national narrative “is one of heroic achievement” (Howard, 2006), the conservative right’s desire to have the curriculum deliver a singular, nationalist narrative has become increasingly more extreme. We risk an “acute crisis of democracy” (Repucci and Slipowitz, 2021, p. 1) as our students are taught a singular narrative that silences First Nations peoples and other cultural minorities. The best defence against this nascent de-democratisation of Australian history classrooms is found in the vital work of history teachers as curriculum workers. If teachers adhere to the curriculum directives focused on historical thinking skills, our students must consider “different perspectives” and use a “range of sources” (ACARA, 2023a) to make evidence-based decisions about our past. The teaching of critical thinking and the use of varied evidence which considers a range of perspectives and assesses their reliability serves as a bulwark against the monocultural assault which seeks to control the content of the curriculum. If we ensure our next generation of citizens have the skills to make informed and critical choices rather than be blind adherents to a nationalist monomyth, our pluralistic liberal democracy will not only survive but thrive.

Developing a composite indicator to assess agricultural sustainability: Influence of some critical choices

Assessing the sustainability of agriculture is a relevant methodological issue with practical implications. In this study, a systematic approach is followed to develop a composite indicator that ranks the sustainability performance of Spanish reference farms at the NUTS 2 level. The classical three-dimension approach is considered under a deductive point of view to represent sustainability. Multicriteria techniques suitable to model the trade-offs between individual attributes are used to weigh and aggregate the three dimensions’ attributes, allowing for different normative standpoints to be expressed. The sensitivity and uncertainty arising from the methodological choices are assessed using a variance-based technique. The ordinary least squares technique is applied to evaluate the sensitivity of the composite indicator to shifts in the consumption of relevant inputs. The results show that the trade-off level does not leverage most analysed units. In addition, the ranking obtained made it possible to discriminate the herbaceous farms as those with the worst sustainability performance. Methodologically, the most influential parameters in the indicator are related to the ethical complexity of evaluating overall sustainability. In addition, the indicator shows a high sensitivity to shifts in the consumption of machinery fuel and irrigation resources. In short, the indicator developed represents a relevant input to improve the interpretation of the sustainability performance of agriculture as a multidimensional issue, emphasising its normative character.

The Customer Perspective on Food Delivery Applications: Insights from Saudi Arabia

In today’s technology-driven era, online platforms and applications have revolutionized various services, including food delivery and offering unparalleled convenience and efficiency. This surge in popularity of digital solutions is driven by consumers' growing technological dependency and their quest for seamless experiences. Consequently, more restaurants outsource delivery services to meet evolving consumer preferences. This study provides business operators with valuable insights into consumer preferences and behaviors regarding online food delivery applications. Specifically, it identifies the determinants influencing the adoption of such apps and explores the most popular apps among the Saudi subpopulation. Additionally, this study investigated the factors associated with the actual usage of online food delivery services. An analysis of the collected data revealed critical factors shaping consumer choices in food delivery services, notably the impact of social media reputation and advertisements. The proximity of restaurants has emerged as a significant determinant, profoundly influencing customer satisfaction and app preferences. Interestingly, while cost did not significantly drive consumers towards online food delivery services, the Jahez app stood out as the preferred choice because of its exemplary customer service. The findings of this study offer valuable insights for informed strategic decision-making within the food delivery ecosystem. Moreover, recommendations for future research are provided to foster a deeper understanding of and innovation in the online food delivery domain.

Enhancing coevolutionary signals in protein–protein interaction prediction through clade-wise alignment integration

Protein–protein interactions (PPIs) play essential roles in most biological processes. The binding interfaces between interacting proteins impose evolutionary constraints that have successfully been employed to predict PPIs from multiple sequence alignments (MSAs). To construct MSAs, critical choices have to be made: how to ensure the reliable identification of orthologs, and how to optimally balance the need for large alignments versus sufficient alignment quality. Here, we propose a divide-and-conquer strategy for MSA generation: instead of building a single, large alignment for each protein, multiple distinct alignments are constructed under distinct clades in the tree of life. Coevolutionary signals are searched separately within these clades, and are only subsequently integrated using machine learning techniques. We find that this strategy markedly improves overall prediction performance, concomitant with better alignment quality. Using the popular DCA algorithm to systematically search pairs of such alignments, a genome-wide all-against-all interaction scan in a bacterial genome is demonstrated. Given the recent successes of AlphaFold in predicting direct PPIs at atomic detail, a discover-and-refine approach is proposed: our method could provide a fast and accurate strategy for pre-screening the entire genome, submitting to AlphaFold only promising interaction candidates—thus reducing false positives as well as computation time.

Impact of Virtual Navigation on the Education and Access of Patients with Cancer: A National Mixed Methods Investigation.

Despite growing evidence of efficacy, oncology patient navigation is not ubiquitously offered. Navigation may reduce barriers to cancer care, yet geographic location may limit patient access. To overcome geographical barriers and increase patient education in oncology, our medical center developed a virtual navigation program. To examine the efficacy of the Virtual Navigation Program designed to increase patient access and education across local and national contexts. In this mixed-methods study, a total of 105 individuals completed an online survey administered nationally to oncology patients who used the navigation program. Clinical and demographic data were collected. Nonparametric tests were used for group comparisons (Wilcoxon Signed-Ranks test). Virtual navigation increased access as the majority of patients utilized the Virtual Navigation Program beyond the local area; 42% of patients were local; 58% were patients located across the US. 55 (52%) were female; 25% were non-White. Pancreatic cancer was the largest tumor type: 51 (49%). Patients agreed the virtual cancer education helped them make critical health choices (mean 4.89 SD = 84); understand their diagnosis (mean 4.44 SD = .77); were able to make informed decisions (mean 4.43 SD = .83), and empowered to manage their cancer care (mean 4.58 SD = .82). Post navigation, the user experience survey revealed significant increases in a subset of patients' cancer knowledge (P < 0.001), access to quality cancer education (P = 0.045), decision-making in cancer care (P < 0.05), and coordinated cancer care (P = 0.03). This study explores a unique role that virtual navigation may play in helping accommodate healthcare for many cancer patients in underserved areas. The use of innovative aspects of research and education in virtual navigation may be successful in promoting patient empowerment in the cancer continuum. Internet-strategies are needed to inform sustainable patient navigation in low-resource contexts.

Development of prognostic model and multivariate analysis for breast cancer survival patients using SEER database

Background: Many studies employed machine learning (ML) to forecast the prognosis of breast cancer (BC) patients and discovered that the ML model showed high individualized forecasting ability. Breast cancer is the most frequent kind of carcinoma in women globally and ranks as the leading cause of death in women. Objectives: This study intends to use the Surveillance, Epidemiology, and End Results dataset to categorize breast carcinoma cases’ alive and dead conditions. Deep learning and machine learning have been extensively utilized in clinical studies to address various categorization problems due to their ability to manage massive data sets in an organized manner. Pre-processing the data allows it to be visualized and analyzed for making critical choices. This study describes a realistic machine learning-based strategy for categorizing the SEER breast cancer dataset. Materials and methods: We employed classification and machine learning algorithms to classify breast cancer mortality. Four well-known classification ML algorithms were employed in this study. To identify risk factors, we employed multivariate analysis using the data set. Results: The decision tree performed the best accuracy (0.914) among all the models. T4 stage (β=1.4, p&lt;0.001, OR=4.22, 95% CI (2.06-8.64), N2 stage (β=0.39, p=0.008, OR= 1.49, 95% CI (1.111-1.997) found to be major risk factors for breast cancer mortality using multivariate analysis. Conclusion: The significant prognostic variables affecting the breast carcinoma survival rates reported in the current research are relevant and might be turned into decision support systems in the medical realm.

Delayed, abrupt and unjust: An institutionalist perspective on limits to climate change adaptation

An actor-centred, risk-based framework for analysing limits to the capacity to adapt to climate change impacts, developed in the context of IPCC AR5 (Dow et al,. 2013b), was refined in the IPCC AR6 reports (Mechler et al., 2020; O’Neill et al., 2022; Thomas et al., 2021). In this paper, we centre the analysis on how institutional contexts shape and influence adaptation limits as experienced by social actors. We emphasise institutions’ stability over time leading to delayed adaptation, their role in protecting more powerful rather than weaker social interests, and their tendency to generate punctuated rather than smooth changes in vulnerability to climate risks. We illustrate these arguments with two case studies of socio-ecological institutional regimes facing adaptation limits: water resources management in the Colorado River Basin; and disaster risk management in The Bahamas. These represent divergent risk governance contexts generating limits to adaptation for multiple social actors, despite the availability of plausible adaptation pathways. Our aim is to contribute to a generalisable approach to adaptation limits which can be applied in identifying and assessing critical choices in responses to growing climate change impacts.

Text-Critical Edition of the List of Solomon’s Officials (3 Kgdms 4:8–19)

Abstract The List of Solomon’s officials (1 Kgs/3 Kgdms 4:8–19) presents many textual problems both in the MT and the LXX texts. It constitutes then a perfect testing ground for the critical edition of 3–4 Kingdoms. This paper intends to revise the Rahlfs’ approach and conclusions, in particular regarding the correspondence εἷς = ‮אֶחָד‬‎ attested differently in the different text forms, and presents a commented preliminary edition of the passage, explaining the critical choices of the editors.

Memory Can Define Individual Beliefs and Identity—and Shape Society

Memory profoundly define individual beliefs and identity, shaping how societies make decisions. Five key memory phenomena include—first impressions and the primacy effect, risky decision-making and memory availability, information reliability and source memory, music preferences and the reminiscence bump, and long-term planning and episodic future thinking. Each phenomenon is explored for its impact on policy, revealing the profound ways that memory biases and processes shape critical societal choices, such as judicial decisions, hiring practices, financial planning, and pro-environmental behavior. Although memory's importance in daily life is widely recognized, its central role in shaping self-identity and influencing societal structures is often underestimated. Memory biases, evident in various aspects, from cultural exposures to decision-making heuristics, play a pivotal role in individual and collective behavior. These biases, often manifesting subtly, guide the formation of beliefs and preferences, with strong implications for policy making. Taken together, this work advocates for a multidisciplinary approach to bridge memory theory and policy practice.

The Use of Trajectories in Management Research: A Review and Insights for Future Research

The increased incorporation of time, theoretically and empirically, has dramatically advanced our understanding of management. Trajectories, which describe longitudinal change in constructs using functions of time, have become one of management scholars’ most widely used methodological tools to examine if, how, and why constructs change. Yet, despite the rapid growth of trajectory research, we lack a coherent framework for understanding how trajectories can be used to generate valuable insights into change-related work phenomena. In this article, we review more than two decades of trajectory research to develop a taxonomy that delineates the theoretical extensions offered by different trajectory-related research questions. In addition, although seminal method articles offered initial how-to instructions for conducting trajectory analyses, there has since been no systematic review of the methodological practices actually used in trajectory research. Our review distinguishes eight critical methodological choices required in every trajectory study, catalogs the frequency of methodological choices made in prior research, identifies problematic practices that persist, and offers actionable recommendations to develop rigorous trajectory research. Finally, we discuss eight pressing issues in current trajectory research, including several prominent practices that must be abandoned, and suggest how to avoid or minimize these concerns in future studies. Together, our review provides management scholars with an organizing platform for developing trajectory-related research questions and a comprehensive guide to rigorous study designs and analyses.

Perception of the Transition to a Zero-Emission Economy in the Opinion of Polish and Foreign Students: A Comparative Analysis

Renewable energy sources serve as an alternative to conventional energy sources, with the main advantages being their inexhaustible nature and the absence of or significant reduction in greenhouse gas emissions. On the other hand, the development of energy production utilising renewable sources remains controversial, primarily because of the high initial costs of RES investment projects and the carbon footprint during the production phase of RES facilities. However, these appear to be gradually offset by the many benefits that outweigh the initial RES facility maintenance costs. Unsurprisingly, European Union (EU) countries are at the forefront of the transition to a low-emission economy. This is evidenced by such things as the 22% decrease in greenhouse gas emissions in the Member States, achieved despite a 58% increase in gross domestic product (GDP) between 1990 and 2017. Nonetheless, the use of renewable energy sources, such as solar, wind, water, etc., is much more expensive than burning fossil fuels. It must be stressed that the European Union’s success in this regard is influenced by the actions taken by all Member States to implement the European Green Deal and their individual targets. Furthermore, the transition to a zero-emission economy is also the result of the informed participation and involvement of all European Community residents. Therefore, there exists a need for education on renewable energy sources to facilitate the transition to a zero-emission economy in European Union countries. This education should primarily target young people, whose critical choices will shape this area in the future. It is the knowledge and awareness of young people that will largely determine whether it will be possible to limit greenhouse gas emissions by 2050. This article aimed to determine how Polish and foreign students view the transition to a zero-emission economy. The article places a special emphasis on renewable energy sources as a key factor in shaping national security in terms of sustainable ecology and friendly social development. The aim of the study was achieved through a research process using theoretical and empirical research methods. To that end, the subject literature and current legislation on renewable energy sources were analysed. Moreover, a survey was conducted among Polish and foreign students living in Poland. The analysis of the research results made it possible to formulate recommendations on how to shape national security in terms of sustainable ecology and friendly social development.

Concept Parser With Multimodal Graph Learning for Video Captioning

Conventional video captioning methods are either stage-wise or simple end-to-end. While the former might introduce additional noise when exploiting off-the-shelf models to provide extra information, the latter suffers from lacking high-level cues. Therefore, a more desired framework should be able to capture multi-aspects of videos consistently. To this end, we present a concept-aware and task-specific model named CAT that accounts for both low-level visual and high-level concept cues, and incorporates them effectively in an end-to-end manner. Specifically, low-level visual and high-level concept features are obtained from the video transformer and concept parser of CAT. And a concept loss is further introduced to regularize the learning process of concept parser w.r.t. generated pseudo ground truth. To combine multi-level features, a caption transformer is later introduced in CAT, where visual and concept features are the inputs and caption is its output. In particular, we make critical design choices in the caption transformer to learn to exploit these cues with a multi-modal graph. This is achieved by a graph loss that enforces effective learning of intra and inter correlations between multi-level cues. Extensive experiments on three benchmark datasets demonstrate that CAT achieves 2.3 and 0.7 improvements in the CIDEr metric on MSVD and MSR-VTT compared to the state-of-the-art method SwinBERT [1], and also achieves a competitive result on VATEX.