Learning Sentence-to-Hashtags Semantic Mapping for Hashtag Recommendation on Microblogs

The growing use of microblogging platforms is generating a huge amount of posts that need effective methods to be classified and searched. In Twitter and other social media platforms, hashtags are exploited by users to facilitate the search, categorization, and spread of posts. Choosing the appropriate hashtags for a post is not always easy for users, and therefore posts are often published without hashtags or with hashtags not well defined. To deal with this issue, we propose a new model, called HASHET ( HAshtag recommendation using Sentence-to-Hashtag Embedding Translation ), aimed at suggesting a relevant set of hashtags for a given post. HASHET is based on two independent latent spaces for embedding the text of a post and the hashtags it contains. A mapping process based on a multi-layer perceptron is then used for learning a translation from the semantic features of the text to the latent representation of its hashtags. We evaluated the effectiveness of two language representation models for sentence embedding and tested different search strategies for semantic expansion, finding out that the combined use of BERT ( Bidirectional Encoder Representation from Transformer ) and a global expansion strategy leads to the best recommendation results. HASHET has been evaluated on two real-world case studies related to the 2016 United States presidential election and COVID-19 pandemic. The results reveal the effectiveness of HASHET in predicting one or more correct hashtags, with an average F -score up to 0.82 and a recommendation hit-rate up to 0.92. Our approach has been compared to the most relevant techniques used in the literature ( generative models , unsupervised models, and attention-based supervised models ) by achieving up to 15% improvement in F -score for the hashtag recommendation task and 9% for the topic discovery task.

Barriers of growth within an informal sector business: narratives of women subsistence entrepreneurs in South Africa

PurposeThe study explores growth barriers experienced by a sample of women subsistence entrepreneurs operating within the informal sector in South Africa.Design/methodology/approachThe paper utilizes a descriptive-exploratory research approach and design relying on semi-structured interviews. A purposive sample of 45 women subsistence entrepreneurs formed the participant pool.FindingsThree main narratives emerged. First, a sense of personal contentment existed as a potential barrier for women subsistence entrepreneurs. Second, the women subsistence entrepreneurs had no expansion strategy due to their circumstances. This served as a barrier to growth. Finally, challenges emanating from the home-front served as a limit to the growth of the informal sector business.Research limitations/implicationsBased on the findings, strategies are offered to assist the women subsistence entrepreneurs in tackling the identified barriers to the growth of the informal sector business. A limitation of the research concerns issues that accompany qualitative research. Notably, these include sampling issues.Practical implicationsBased on the findings, strategies are offered to assist women subsistence entrepreneurs in tackling the barriers that affect their businesses.Originality/valueGiven the popularity of the informal sector in emerging nations such as South Africa, the study proffers suggestions that assist the advancement of subsistence entrepreneurship, especially within the informal sector. The role of women in all this is heightened.

Applications of a geographic information system–based business intelligence system for decision making at Coffee Refresh

This case is designed to illustrate how to utilize a business intelligence framework and a geographic information system to make better decisions on franchise opening requests. The case started when Jong, the senior certified analytics professional, and his team were drafting a proposal presentation on a new franchise approval process to present to Tony, the director of Coffee Refresh. Currently, Coffee Refresh was experiencing a significant delay in approving franchise opening requests. In addition, the current approval process was relied on the appraisal teams’ evaluations, which involved self-judgment as to the accessibility and visibility of the store location in each franchise opening request and a manual count of the potential targeted customers at each location, which impacted how the appraisal teams estimated the expected revenue for each location. During the first meeting with Jong and his team, Tony was frustrated by the current evaluation process, resulting in a significant backlog of requests pending approval, which was far behind the target for its franchise branch expansion strategy. It had also been reported that the branches that opened during the past few years had a relatively low success ratio (losses between 4.8% and 16.5%) and that the rate of branch closures had increased from 2% to 7.5%. Currently, eight franchise opening requests were pending, and Jong had to provide a recommendation for which requests should be approved. Coffee Refresh could approve all of them instantly, approve some of them, or even decline all of them. This was a great opportunity for Jong and his team to revisit the current franchise approval process and to demonstrate how business intelligence systems could improve the request approval process as well as address the issues of the current franchises’ underperformance and the closure rates.

Fast-RRT: A RRT-Based Optimal Path Finding Method

As a sampling-based pathfinding algorithm, Rapidly Exploring Random Trees (RRT) has been widely used in motion planning problems due to the ability to find a feasible path quickly. However, the RRT algorithm still has several shortcomings, such as the large variance in the search time, poor performance in narrow channel scenarios, and being far from the optimal path. In this paper, we propose a new RRT-based path find algorithm, Fast-RRT, to find a near-optimal path quickly. The Fast-RRT algorithm consists of two modules, including Improved RRT and Fast-Optimal. The former is aims to quickly and stably find an initial path, and the latter is to merge multiple initial paths to obtain a near-optimal path. Compared with the RRT algorithm, Fast-RRT shows the following improvements: (1) A Fast-Sampling strategy that only samples in the unreached space of the random tree was introduced to improve the search speed and algorithm stability; (2) A Random Steering strategy expansion strategy was proposed to solve the problem of poor performance in narrow channel scenarios; (3) By fusion and adjustment of paths, a near-optimal path can be faster found by Fast-RRT, 20 times faster than the RRT* algorithm. Owing to these merits, our proposed Fast-RRT outperforms RRT and RRT* in both speed and stability during experiments.

ONE STOP KITCHEN: managing a multi-brand cloud kitchen

Subject area Marketing; Business Expansion Strategy Case synopsis Rahul Hazarika, the owner of One Stop Kitchen (OSK), was convinced that expansion of his cloud kitchen had to be aligned at various fronts to compete. He was juggling with thoughts of expanding from B2C to B2B model, whether to expand in the same city or other cities or to introduce new variants in existing B2C so as to use and leverage the existing customer base and network. The competition was growing ablaze, and most of the orders were procured from the third-party aggregators, leaving little chance for OSK to connect to its customer base directly. The growth potential in this upcoming sector was undeniable, and Hazarika was continually trying to bring in relevant business practices to maintain the momentum. The case deals growing cloud kitchen business problems and covers business concepts such as business expansion strategy, product profiling and marketing strategies. Learning objectives Students will understand strategies for a Online startup Business in the Cloud Kitchen category. The concept of Product Profiling is also dealt with in the case. Complexity academic level Suitable for PG- and Executive-level courses. Supplementary materials Teaching notes are available for educators only. Subject code CSS 8: Marketing.

Selective YAP activation in Procr cells is essential for ovarian stem/progenitor expansion and epithelium repair

Ovarian surface epithelium (OSE) undergoes recurring ovulatory rupture and OSE stem cells rapidly generate new cells for the repair. How the stem cell senses the rupture and promptly turns on proliferation is unclear. Our previous study has identified that Protein C Receptor (Procr) marks OSE progenitors. In this study, we observed decreased adherent junction and selective activation of YAP signaling in Procr progenitors at OSE rupture site. OSE repair is impeded upon deletion of Yap in these progenitors. Interestingly, Procr+ progenitors show lower expression of Vgll4, an antagonist of YAP signaling. Overexpression of Vgll4 in Procr+ cells hampers OSE repair and progenitor proliferation, indicating that selective low Vgll4 expression in Procr+ progenitors is critical for OSE repair. In addition, YAP activation promotes transcription of the OSE stemness gene Procr. The combination of increased cell division and Procr expression leads to expansion of Procr+ progenitors surrounding the rupture site. These results illustrate a YAP-dependent mechanism by which the stem/progenitor cells recognize the ovulatory rupture, and rapidly multiply their numbers, highlighting a YAP-induced stem cell expansion strategy.

Mobilized Mouse and Human Peripheral Blood Containing Elevated Numbers of Donor Treg Cells Ameliorates Pre-Clinical GvHD and GVL Is Maintained in an MHC-Matched Allogeneic Murine Model

Mobilized hematopoietic stem / progenitor cells have become the primary option used for adult allogeneic hematopoietic stem cell transplants (aHSCTs) although GVHD remains a major immunologic complication (CIBMTR, 2017). The present studies are directed to developing a translational model for the application of donor regulatory T cells (Tregs) together with the use of mobilized peripheral blood (mPB) for more widespread application of these aHSCTs. We have previously reported that two-pathway Treg expansion strategy markedly elevates Treg numbers and function (Wolf, 2017; Copsel 2018). Additionally, this approach could be effectively applied in a concomitant strategy to amplify Treg suppressor activity in mPB to diminish pre-clinical GVHD (Barreras H, ASTCT Meeting S292 ,2019). Here, we present new findings using mPB transplants examining GVL reactivity in an MHC-matched minor antigen mismatched pre-clinical model and test the notion that human donor Tregs can be employed with mobilized healthy donor PB to effectively regulate xenogeneic GVHD. Donor C3H.SW (H2b) mice were treated with filgrastim for 4 days prior to use for transplants and PB analysis demonstrated significant elevations in c-kit + cells. Some PB donors were concurrently administered the two-pathway Treg expansion strategy (D-6 through D-1) consisting of TL1A-Ig, targeting TNFRSF25, and low dose huIL-2. A significant increase (up to 40%) in the frequency of CD4 +FoxP3 + Tregs occurred during the mobilization process. Pooled C3H.SW PB was collected from both mobilized and Treg unexpanded ("TrUM") or mobilized and Treg expanded ("TrEM") donors and transplanted into lethally irradiated, MHC-matched B6 (H2b) recipients. Recipients of TrEM donors exhibited significantly reduced weight loss and clinical GVHD scores compared to recipients of TrUM (Fig.1A). GVL responses were tested in animals administered B6-MLL-AF9 GFP leukemia cells administered at the time of transplant. Notably, recipients of TrEM exhibited comparable GVL activity to TrUM recipients evidenced by B6-MLL-AF9 GFP levels in bone marrow and spleen (Fig.1B). To begin to translate these findings, we tested the use of ex-vivo expanded human donor Tregs (huTregs) to ameliorate xenogeneic GVHD (xGVHD). First, sorted huTregs (CD4 +CD25 +CD127 lo) from a healthy donor were expanded ex-vivo using anti-CD3/anti-CD28 beads 1 week prior to transplant (Fig 2A). next, these expanded huTregs were combined with human mPB from the same healthy donor (6x10 6 PBMC) and transplanted into NSG/ NOD-scid IL2Rgamma null mice. We found that treatment with ex-vivo expanded huTregs resulted in significant reduction of mortality rate and clinical xGVHD (Fig 2B,C). Notably, 1 week post-transplant PB huTregs levels were still elevated and frequency of huCD4 +Tconv cells was diminished supporting xGVHD outcomes (Fig 2D). In total, these findings demonstrated that the use of mPB containing elevated Treg levels significantly reduced pre-clinical GVHD without loss of GVL activity in an MHC-matched allogeneic model. Moreover, utilizing ex-vivo expanded huTregs from a mPB donor and added back to the same donor at the time of transplant ameliorated xGVHD. The observations indicate that during the donor stem / progenitor cell mobilization process, manipulation of donor Tregs using our two-pathway strategy can be successfully accomplished in PB resulting in an effective and translational approach to ameliorate GVHD following aHSCT. In total, the present studies support the notion that in vivo or ex-vivo manipulation of donor tregs together with mobilized peripheral blood could provide therapeutic approaches to improve aHSCT outcomes. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.