This thesis develops a trainable object-recognition algorithm. This algorithm represents objects using their salient features. The algorithm applies an attention mechanism to speed up feature detection. A trainable component-based object recognition system which implements the developed algorithm has been created. This system has two layers. The first layer contains several individual feature classifiers. They detect salient features which compose higher level objects from input images. The second layer judges if those detected features form a valid object. An object is represented by a feature map which stores the geometrical and hierarchical relations among features and higher level objects. It is the input to the second layer. The attention mechanism is applied to improve feature detection speed. This mechanism will lead the system to areas with a higher likelihood of containing features when a few features are detected. Therefore the feature detection will be sped up. Two major experiments are conducted. These experiments applied the de- veloped system to discriminate faces from non-faces and to discriminate people from backgrounds in thermal images. The results of these experiments show the success of the implemented system. The attention mechanism displays a positive effect on feature detection. It can save feature detection time, especially in terms of classifier calls.

Purpose: To explore the influence of marketing AI on marketing managers' attention allocation to leverage customer relationships in different business-to-business contexts. Method: Abductive approach and multiple case study, data collection was made by qualitative semi-structured interviews and secondary data collection. Conclusion: The study identified both similarities and differences within three main categories across two industrial clusters where marketing AI effect marketing managers’ attention allocation in B2B relationships. Marketing AI is shown to affect B2B relationships through marketing managers’ attentional selection towards efficiencies and/or new opportunities. Marketing AI is shown to influence marketing managers’ attention allocation by distorting the focus of attention on relational dynamics by introducing automated or augmented marketing AI solutions into the relationship.

Image captioning is a crucial field across computer vision and natural language processing. It could be widely applied to high-volume web images, such as conveying image content to visually impaired users. Many methods are adopted in this area such as attention-based methods, semantic-concept based models. These achieve excellent performance on general image datasets such as the MS COCO dataset. However, it is still left unexplored on single-object images.In this paper, we propose a new attribute-information-combined attention- based network (AIC-AB Net). At each time step, attribute information is added as a supplementary of visual information. For sequential word generation, spatial attention determines specific regions of images to pass the decoder. The sentinel gate decides whether to attend to the image or to the visual sentinel (what the decoder already knows, including the attribute information). Text attribute information is synchronously fed in to help image recognition and reduce uncertainty.We build a new fashion dataset consisting of fashion images to establish a benchmark for single-object images. This fashion dataset consists of 144,422 images from 24,649 fashion products, with one description sentence for each image. Our method is tested on the MS COCO dataset and the proposed Fashion dataset. The results show the superior performance of the proposed model on both multi-object images and single-object images. Our AIC-AB net outperforms the state-of-the-art network, Adaptive Attention Network by 0.017, 0.095, and 0.095 (CIDEr Score) on the COCO dataset, Fashion dataset (Bestsellers), and Fashion dataset (all vendors), respectively. The results also reveal the complement of attention architecture and attribute information. / Bildtextning är ett avgörande fält för datorsyn och behandling av naturligt språk. Det kan tillämpas i stor utsträckning på högvolyms webbbilder, som att överföra bildinnehåll till synskadade användare. Många metoder antas inom detta område såsom uppmärksamhetsbaserade metoder, semantiska konceptbaserade modeller. Dessa uppnår utmärkt prestanda på allmänna bilddatamängder som MS COCO-dataset. Det lämnas dock fortfarande outforskat på bilder med ett objekt.I denna uppsats föreslår vi ett nytt attribut-information-kombinerat uppmärksamhetsbaserat nätverk (AIC-AB Net). I varje tidsteg läggs attributinformation till som ett komplement till visuell information. För sekventiell ordgenerering bestämmer rumslig uppmärksamhet specifika regioner av bilder som ska passera avkodaren. Sentinelgrinden bestämmer om den ska ta hand om bilden eller den visuella vaktposten (vad avkodaren redan vet, inklusive attributinformation). Text attributinformation matas synkront för att hjälpa bildigenkänning och minska osäkerheten.Vi bygger en ny modedataset bestående av modebilder för att skapa ett riktmärke för bilder med en objekt. Denna modedataset består av 144 422 bilder från 24 649 modeprodukter, med en beskrivningsmening för varje bild. Vår metod testas på MS COCO dataset och den föreslagna Fashion dataset. Resultaten visar den överlägsna prestandan hos den föreslagna modellen på både bilder med flera objekt och enbildsbilder. Vårt AIC-AB-nät överträffar det senaste nätverket Adaptive Attention Network med 0,017, 0,095 och 0,095 (CIDEr Score) i COCO-datasetet, modedataset (bästsäljare) respektive modedatasetet (alla leverantörer). Resultaten avslöjar också komplementet till uppmärksamhetsarkitektur och attributinformation.

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第23311号 / 情博第747号 / 新制||情||128(附属図書館) / 京都大学大学院情報学研究科知能情報学専攻 / (主査)准教授 吉井 和佳, 教授 河原 達也, 教授 西野 恒, 教授 鹿島 久嗣 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Machine translation (MT) as a field of research has known significant advances in recent years, with the increased interest for neural machine translation (NMT). By combining deep learning with translation, researchers have been able to deliver systems that perform better than most, if not all, of their predecessors. While the general consensus regarding NMT is that it renders higher-quality translations that are overall more idiomatic, researchers recognize that NMT systems still struggle to deal with certain classic difficulties, and that their performance may vary depending on their architecture. In this project, we implement a challenge-set based approach to the evaluation of examples of three main NMT architectures: convolutional neural network-based systems (CNN), recurrent neural network-based (RNN) systems, and attention-based systems, trained on the same data set for English to French translation. The challenge set focuses on a selection of lexical and syntactic difficulties (e.g., ambiguities) drawn from literature on human translation, machine translation, and writing for translation, and also includes variations in sentence lengths and structures that are recognized as sources of difficulties even for NMT systems. This set allows us to evaluate performance in multiple areas of difficulty for the systems overall, as well as to evaluate any differences between architectures’ performance. Through our challenge set, we found that our CNN-based system tends to reword sentences, sometimes shifting their meaning, while our RNN-based system seems to perform better when provided with a larger context, and our attention-based system seems to struggle the longer a sentence becomes.

With the recent advantages of machine learning in cheminformatics, the drug discovery process has been accelerated; providing a high impact in the field of medicine and public health. Molecular property and activity prediction are key elements in the early stages of drug discovery by helping prioritize the experiments and reduce the experimental work. In this thesis, a novel approach for multi-task regression using a text-based Transformer model is introduced and thoroughly explored for training on a number of properties or activities simultaneously. This multi-task regression with Transformer based model is inspired by the field of Natural Language Processing (NLP) which uses prefix tokens to distinguish between each task. In order to investigate our architecture two data categories are used; 133 biological activities from ExCAPE database and three physical chemistry properties from MoleculeNet benchmark datasets. The Transformer model consists of the embedding layer with positional encoding, a number of encoder layers, and a Feedforward Neural Network (FNN) to turn it into a regression problem. The molecules are represented as a string of characters using the Simplified Molecular-Input Line-Entry System (SMILES) which is a ’chemistry language’ with its own syntax. In addition, the effect of Transfer Learning is explored by experimenting with two pretrained Transformer models, pretrained on 1.5 million and on 100 million molecules. The text-base Transformer models are compared with a feature-based Support Vector Regression (SVR) with the Tanimoto kernel where the input molecules are encoded as Extended Connectivity Fingerprint (ECFP), which are calculated features. The results have shown that Transfer Learning is crucial for improving the performance on both property and activity predictions. On bioactivity tasks, the larger pretrained Transformer on 100 million molecules achieved comparable performance to the feature-based SVR model; however, overall SVR performed better on the majority of the bioactivity tasks. On the other hand, on physicochemistry property tasks, the larger pretrained Transformer outperformed SVR on all three tasks. Concluding, the multi-task regression architecture with the prefix token had comparable performance with the traditional feature-based approach on predicting different molecular properties or activities. Lastly, using the larger pretrained models trained on a wide chemical space can play a key role in improving the performance of Transformer models on these tasks.

Due to the importance of corporate entrepreneurship and leadership innovation when nurturing and building the company, along with keeping up with competition in an ever developing world these concepts have become more of an importance over the past few years. Corporate Entrepreneurship has an important role in building innovation and is needed to avoid disruption. However, the level of conscious attention given to CE within Swedish SMEs is still to some extent unknown. This study aims to investigate the CE initiatives produced and the active awareness put towards it along with the extent entrepreneurial leadership plays a role in the organizational innovation. The two research questions are: How are corporate leaders paying attention to corporate entrepreneurship? & How is entrepreneurial leadership enhancing leadership innovation in Swedish SMEs? This study focuses on Swedish small to medium-sized companies of various sizes and within different industries but also acting within the same ownership structure. For this reason, the relationship and entrepreneurial culture within the whole organization is also of interest. Therefore, a total of four interviews with six individuals from four different companies were conducted. The data was analyzed using a thematic analysis method. The results showed that there is an unconscious use of CE approaches and initiative along with the important role of entrepreneurial leadership to maintain and develop the organization.

[pt] Diante da complexidade do ambiente de negócios contemporâneo e tendo em vista as limitações da racionalidade humana, os estudos da teoria do comportamento organizacional propostos e a teoria da visão baseada na atenção, este estudo objetiva diagnosticar o nível de atenção a temas críticos para a proteção empresarial e verificar a percepção sobre o nível de risco de situações propostas, por meio de pesquisa aplicada a profissionais que desempenham função gratificada em uma organização do setor de óleo e gás. Para tanto, com base em estudos desenvolvidos de atenção e percepção de risco, foi realizada pesquisa quantitativa de natureza descritiva, respondida voluntariamente por 409 profissionais ocupantes de função gratificada de diferentes áreas da empresa: de negócios e corporativas, de diferentes níveis organizacionais, níveis de formação, tempo de empresa e idade. Os resultados apontam que os temas relacionados à logística de produtos e pessoas são os assuntos com maior engajamento na captação da atenção, análise dos riscos relacionados e sobre os quais os profissionais participantes desta pesquisa investem mais tempo e esforço na busca de mais informações/conhecimentos. / [en] Given the complexity of the contemporary business environment and considering the limitations of human rationality, the proposed studies of the organizational behavior theory and the vision based in attention theory, this study aims to diagnose the level of attention to critical issues for corporate security and verify and the perception of the risk level of proposed situations, through research applied to professionals who perform a gratified position in an organization in the oil and gas sector. Therefore, based on studies developed on attention and risk perception, a quantitative descriptive research was carried out, answered voluntarily by 409 professionals occupying a gratified position in different areas on the company: business and the corporate, from different organizational levels, education level, time in the organization and age. The results show that topics related to the topics of products and people are the subjects with the greatest engagement in attracting attention, analyzing the related risks and on which the professionals participating in this research invest more time and the search for more information/knowledge.

This thesis examines the decision-making regarding market-orientation within specific types of Swedish state-owned enterprises, namely the Public Business Authorities (Affärsverk, PBAs), during a period of institutional upheaval in the 1980s. It specifically focuses on the leadership groups – director general and board – of the Energy PBA (Statens Vattenfallsverk), the Postal PBA (Postverket), and the Tele PBA (Televerket). The thesis adopts a theoretical framework that integrates the perspectives of varieties of capitalism and the attention-based view of the firm, which posits that firm behaviour is determined by where and how attention is directed within the organisation. The thesis delineates two distinct leadership strategies for addressing market-orientation: the deliberative institutional change strategy, characterised by consensus-building through negotiation using existing modes of strategic interaction, and the entrepreneurial institutional change strategy, characterised by leveraging existing modes of strategic interaction to transcend them and forge novel paths towards new modes of strategic interaction. The former was applied by the Postal and Energy PBAs, while the latter was applied by the Tele PBA.

Within dynamic 21st century knowledge economies, future-building knowledge, that bears capacities to transcend existing boundaries and create something new, is of particular importance. Within the first decade of the new century, new concepts such as "learning from the future" or "self-transcending knowledge" developed within knowledge management. So far, they lacked a theoretical grounding in relevant learning theory as well as a sound acknowledgement and consideration of such knowledge structures' emergence and social embeddedness. Thus, key principles and leverage factors for designing respective knowledge processes were difficult to derive. This dissertation investigates theoretical ground that can provide a basis to explain the creation of future-building knowledge in collective structures. It is guided by the following research question: "How can the emergence of self-transcending knowledge in collective organizational settings be rooted in theories of knowledge, learning and cognition?" Starting from the model of knowledge-based management, the model is expanded by exploring cognitive, creative and social systemic aspects of knowledge creation on a transdisciplinary basis. Research draws on constructivist learning theory, complexity-based approaches in knowledge management and organizational learning, recent accounts in cognitive science (enaction/embodiment) and a creative logic of emergence to derive an integrated model for collective self-transcending knowledge. The model contributes to the integration of knowledge management, organizational learning and cognitive science, expanding knowledge-based management towards attention-based management. The model's three dimensions and three domains form an integrated theoretical basis to derive key principles and leverage factors for steering future-building knowledge processes. Simultaneously, they reveal leverage factors' limited - i.e. enabling, not determining - impact on processes of "organizing future". (author's abstract)