Syftet med den här uppsatsen är att analysera läromedel som är publicerat på lektion.se och som har direkt anknytning till litteraturundervisning. Jag gör en kvalitativ textanalys av materialet och undersöker vilka frågor materialet ställer om litteraturen och hur uppgifterna är formulerade, för att därigenom undersöka vilken undervisningspraktik materialet speglar. I min teoretiska tolkning använder jag tre perspektiv på tolkning av text: sändarcentrerat perspektiv, textcentrerat perspektiv och läsar- och erfarenhetsbaserat perspektiv. Två andra centrala teoretiska begrepp är litterär kompetens och matchningsteknik. Jag kommer fram till att mycket av materialet är utformat med kontrollfrågor till de litterära texter som frågorna och uppgifterna riktar sig till. Detta kan visa på att undervisningspraktiken till stor del är monologisk utformad där läraren är den som har tolkningsföreträde till texterna och också den som har de ”rätta svaren”. När uppgifterna istället är autentiskt ställda, så att de möjliggör för en dialogiskt utformad undervisning, så pekar de ofta bort från texten och kräver väldigt lite i form av tolkning, kritiskt tänkande och reflektion. De autentiska frågorna ställs ofta på bekostnad av de texter som de ämnar analysera.

In this paper, we propose a novel active learning algorithm for short-text (Chinese) classification applied to a deep learning architecture. This topic thus belongs to a cross research area between active learning and deep learning. One of the bottlenecks of deeplearning for classification is that it relies on large number of labeled samples, which is expensive and time consuming to obtain. Active learning aims to overcome this disadvantage through asking the most useful queries in the form of unlabeled samples to belabeled. In other words, active learning intends to achieve precise classification accuracy using as few labeled samples as possible. Such ideas have been investigated in conventional machine learning algorithms, such as support vector machine (SVM) for imageclassification, and in deep neural networks, including convolutional neural networks (CNN) and deep belief networks (DBN) for image classification. Yet the research on combining active learning with recurrent neural networks (RNNs) for short-text classificationis rare. We demonstrate results for short-text classification on datasets from Zhuiyi Inc. Importantly, to achieve better classification accuracy with less computational overhead,the proposed algorithm shows large reductions in the number of labeled training samples compared to random sampling. Moreover, the proposed algorithm is a little bit better than the conventional sampling method, uncertainty sampling. The proposed activelearning algorithm dramatically decreases the amount of labeled samples without significantly influencing the test classification accuracy of the original RNNs classifier, trainedon the whole data set. In some cases, the proposed algorithm even achieves better classification accuracy than the original RNNs classifier. / I detta arbete studerar vi en ny aktiv inlärningsalgoritm som appliceras på en djup inlärningsarkitektur för klassificering av korta (kinesiska) texter. Ämnesområdet hör därmedtill ett ämnesöverskridande område mellan aktiv inlärning och inlärning i djupa nätverk .En av flaskhalsarna i djupa nätverk när de används för klassificering är att de beror avtillgången på många klassificerade datapunkter. Dessa är dyra och tidskrävande att skapa. Aktiv inlärning syftar till att överkomma denna typ av nackdel genom att generera frågor rörande de mest informativa oklassade datapunkterna och få dessa klassificerade. Aktiv inlärning syftar med andra ord till att uppnå bästa klassificeringsprestanda medanvändandet av så få klassificerade datapunkter som möjligt. Denna idé har studeratsinom konventionell maskininlärning, som tex supportvektormaskinen (SVM) för bildklassificering samt inom djupa neuronnätverk inkluderande bl.a. convolutional networks(CNN) och djupa beliefnetworks (DBN) för bildklassificering. Emellertid är kombinationenav aktiv inlärning och rekurrenta nätverk (RNNs) för klassificering av korta textersällsynt. Vi demonstrerar här resultat för klassificering av korta texter ur en databas frånZhuiyi Inc. Att notera är att för att uppnå bättre klassificeringsnoggranhet med lägre beräkningsarbete (overhead) så uppvisar den föreslagna algoritmen stora minskningar i detantal klassificerade träningspunkter som behövs jämfört med användandet av slumpvisadatapunkter. Vidare, den föreslagna algoritmen är något bättre än den konventionellaurvalsmetoden, osäkherhetsurval (uncertanty sampling). Den föreslagna aktiva inlärningsalgoritmen minska dramatiskt den mängd klassificerade datapunkter utan att signifikant påverka klassificeringsnoggranheten hos den ursprungliga RNN-klassificeraren när den tränats på hela datamängden. För några fall uppnår den föreslagna algoritmen t.o.m.bättre klassificeringsnoggranhet än denna ursprungliga RNN-klassificerare.

Chatbots are the next big improvement in the era of conversational services. A chatbot is a virtual person who can carry out a conversation with a human about a certain subject, using interactive textual skills. Currently, there are many cloud-based chatbots services that are being developed and improved such as IBM Watson, well known for winning the quiz show “Jeopardy!” in 2011. Chatbots are based on a large amount of structured data. They contains many examples of questions that are associated to a specific intent which represents what the user wants to say. Those associations are currently being done by hand, and this project focuses on improving this data structuring using both supervised and unsupervised algorithms. A supervised reclassification using an improved Barycenter method reached 85% in precision and 75% in recall for a data set containing 2005 questions. Questions that did not match any intent were then clustered in an unsupervised way using a K-means algorithm that reached a purity of 0.5 for the optimal K chosen. / Chatbots är nästa stora förbättring i konversationstiden. En chatbot är en virtuell person som kan genomföra en konversation med en människa om ett visst ämne, med hjälp av interaktiva textkunskaper. För närvarande finns det många molnbaserade chatbots-tjänster som utvecklas och förbättras som IBM Watson, känt för att vinna quizshowen "Jeopardy!" 2011. Chatbots baseras på en stor mängd strukturerade data. De innehåller många exempel på frågor som är kopplade till en specifik avsikt som representerar vad användaren vill säga. Dessa föreningar görs för närvarande för hand, och detta projekt fokuserar på att förbättra denna datastrukturering med hjälp av både övervakade och oövervakade algoritmer. En övervakad omklassificering med hjälp av en förbättrad Barycenter-metod uppnådde 85 % i precision och 75 % i recall för en dataset innehållande 2005 frågorna. Frågorna som inte matchade någon avsikt blev sedan grupperade på ett oövervakad sätt med en K-medelalgoritm som nådde en renhet på 0,5 för den optimala K som valts.

In multimedia learning environments, research suggests that simultaneous presentation of redundant text (i.e. identical narration and on-screen text) may inhibit learning when presented with animation at the same time. However, related studies are limited to testing with cause-and-effects content information (e.g., Moreno & Mayer, 1999, 2002). This study examined the effects of redundant text on learners' memory achievement and problem solving ability. The study replicated and extended prior research by using descriptive, rather than cause-and-effect content information. The primary research questions were (a) does redundant text improve learning performance if learners are presented with instructional material that addresses subject matter other than cause-and-effect relationship? and (b) does sequential presentation of animation followed by redundant text help learning? To answer the research questions, five hypotheses were tested with a sample of 224 Taiwanese students enrolled in a college level Management Information System (MIS) courses at a management college in southern Taiwan. Statistically significant differences were found in memory achievement and problem solving test scores between simultaneous and sequential groups; while no statistically significant differences were found in memory achievement and problem solving test scores between verbal redundant and non-redundant groups. These results were supported by interviewees expressing difficulty in connecting animation and verbal explanation in the two sequential presentation groups. The interview responses also helped to explain why insignificant results were obtained when redundant and non-redundant verbal explanations with animation were presented simultaneously. In general, the results support previous research on the contiguity principle, suggesting that sequential presentations may lead to lower learning performance when animation and verbal explanation are closely related. The separation of the two types of information may increase cognitive load. In addition, the study found that impairment of redundant text was also affected by various learning characteristics, such as the structure of the instructional content and learners previous learning experiences. Recommendations for future study include: (a) research on various situations such as characteristics of the content, characteristics of learners, and difficulty of the instructional material that influences the effects of redundant text, and (b) research on prior learning experience that influences the effects of simultaneous redundant text presentations.

Artificial intelligence (AI) has made astonishing breakthroughs in recent years and achieved comparable or even better performance compared to humans on many real-world tasks and applications. However, it is still far from reaching human-level intelligence in many ways. Specifically, although AI may take inspiration from neuroscience and cognitive psychology, it is dramatically different from humans in both what it learns and how it learns. Given that current AI cannot learn as effectively and efficiently as humans do, a natural solution is analyzing human learning processes and projecting them into AI design. This dissertation presents three studies that examined cognitive theories and established frameworks to integrate crucial human cognitive learning elements into AI algorithms to build human learning–augmented AI in the context of text analytics. The first study examined compositionality—how information is decomposed into small pieces, which are then recomposed to generate larger pieces of information. Compositionality is considered as a fundamental cognitive process, and also one of the best explanations for humans' quick learning abilities. Thus, integrating compositionality, which AI has not yet mastered, could potentially improve its learning performance. By focusing on text analytics, we first examined three levels of compositionality that can be captured in language. We then adopted design science paradigms to integrate these three types of compositionality into a deep learning model to build a unified learning framework. Lastly, we extensively evaluated the design on a series of text analytics tasks and confirmed its superiority in improving AI's learning effectiveness and efficiency. The second study focused on transfer learning, a core process in human learning. People can efficiently and effectively use knowledge learned previously to solve new problems. Although transfer learning has been extensively studied in AI research and is often a standard procedure in building machine learning models, existing techniques are not able to transfer knowledge as effectively and efficiently as humans. To solve this problem, we first drew on the theory of transfer learning to analyze the human transfer learning process and identify the key elements that elude AI. Then, following the design science paradigm, a novel transfer learning framework was proposed to explicitly capture these cognitive elements. Finally, we assessed the design artifact's capability to improve transfer learning performance and validated that our proposed framework outperforms state-of-the-art approaches on a broad set of text analytics tasks. The two studies above researched knowledge composition and knowledge transfer, while the third study directly addressed knowledge itself by focusing on knowledge structure, retrieval, and utilization processes. We identified that despite the great progress achieved by current knowledge-aware AI algorithms, they are not dealing with complex knowledge in a way that is consistent with how humans manage knowledge. Grounded in schema theory, we proposed a new design framework to enable AI-based text analytics algorithms to retrieve and utilize knowledge in a more human-like way. We confirmed that our framework outperformed current knowledge-based algorithms by large margins with strong robustness. In addition, we evaluated more intricately the efficacy of each of the key design elements. / Doctor of Philosophy / This dissertation presents three studies that examined cognitive theories and established frameworks to integrate crucial human cognitive learning elements into artificial intelligence (AI) algorithm designs to build human learning–augmented AI in the context of text analytics. The first study examined compositionality—how information is decomposed into small pieces, which are then recomposed to generate larger pieces of information. Design science research methodology has been adopted to propose a novel deep learning–based framework that can incorporate three levels of compositionality in language with significantly improved learning performance on a series of text analytics tasks. The second study went beyond that basic element and focused on transfer learning—how humans can efficiently and effectively use knowledge learned previously to solve new problems. Our novel transfer learning framework, which is grounded in the theory of transfer learning, has been validated on a broad set of text analytics tasks with improved learning effectiveness and efficiency. Finally, the third study directly addressed knowledge itself by focusing on knowledge structure, retrieval, and utilization processes. We drew on schema theory and proposed a new design framework to enable AI-based text analytics algorithms to retrieve and utilize knowledge in a more human-like way. Lastly, we confirmed our design's superiority in dealing with knowledge on several common text analytics tasks compared to existing knowledge-based algorithms.

This dissertation explores "fairness" as an ethical construct within federal procurement contracting using 3,548 contract dispute decisions published by the Armed Services Board of Contract Appeals (ASBCA) between 2007 and 2021. It employed a multi-faceted, mixed method research design at macro, mezzo, and micro levels that used a blend of descriptive analysis, computational text analysis, and qualitative thematic analysis to explore a little-studied operational domain within public administration. This investigative approach made possible an examination of how fairness manifests in federal procurement in three aspects: equality (competition), equity (contractor demographic identity), and reciprocity (dispute resolution outcome). Aspects of Moore's Public Values Framework were combined with Lipsky's theories regarding street-level bureaucracy and Maynard-Moody and Musheno's conceptualization of frontline workers as knowledge agents to examine the "human" dimensions of administrative discretion in procurement. In addition to explaining the fundamental differences between "fairness" (between individual entities) and "justice" (fairness writ large at the societal level), the dissertation demonstrates how power dynamics between the government sovereign and its commercial civilian partners complicate contract relationships. Its quantitative findings suggest that fairness is impacted by procurement complexity, entrenched arms-length contracting relationships, and strictly construed risk apportionment when contingencies adversely impact contract performance conditions, and that contractor identity plays some role (though its extent is unclear) in the generation and resolution of particularly contentious disputes. This study's qualitative findings indicate that both parties perceive a breakdown in the contractual duty of "good faith and fair dealing" when rivalry is pursued over cooperation, when the parties fail to understand or respect each other's responsibilities and constraints, and when the behavior of government contracting officials creates role confusion between the protection of government interests and the legislatively required fair treatment of contracting partners. Ultimately, this dissertation speaks to ongoing discussions in diverse fields and disciplines such as public administration, organizational studies, empirical legal research, and relational contracting. It also contributes to developing theories regarding complexity in procurement and existing contracting studies from both sociological and economic perspectives. / Doctor of Philosophy / In popular thought, written contracts exist to protect the rights of both parties should one fail to uphold its part of the "bargain." Some legal theorists argue, by contrast, that the contracting process fundamentally is about interpersonal relationships, and that litigated contract disputes are not merely about material redress, but moreover, a failure of the "spirit of contract." From this perspective, a contract's true value lies more in the quality of the relationships it creates than in its documentary perfection. Interpersonal fairness, where the parties treat each other and their contract promises with integrity and respect, is a key component of that relationship. This dissertation studied the ethical expectation of "fairness" in federal defense contracts using 3,548 formal contract dispute decisions published by the Armed Services Board of Contract Appeals (ASBCA) between 2007 and 2021. These decisions were used to examine what procurement fairness means by focusing on three aspects: equality (fair competition for business opportunities), equity (fair distribution of public funding), and reciprocity (how the "spirit of contract" is honored during contract administration). The study explored how government sovereignty impacts contractors' expectations of fair treatment. It further demonstrated that contract relationships are challenged by the complex technical, administrative, and legal requirements of federal contracts. The study's findings revealed that the most contentious disputes (those that require a judge's ruling on legal merit) result from highly competitive contracts where maximum risk has been placed on contractors for performance and price control. The findings also suggested that contractor demographic identity plays some part in how disputes begin and how they are resolved, though the extent and implications of these differences are unclear. Finally, the study indicated that disputes alleging a violation of the contractual duty of "good faith and fair dealing" showed evidence of entrenched rivalry instead of cooperation, the contracting parties' failure to appreciate each other's operating challenges and constraints, and confusion about how federal contracting officers function as both protectors of the government's interests and as contractor rights advocates under federal contract law.

Text-to-speech (TTS) systems are ubiquitous. From Siri to Alexa to customer service phone call options, listening in a real-world context requires language learners to interact with TTS. Traditionally, language learners report difficulty when listening due to various reasons including genre, text, task, speaker characteristics, and environmental factors. This naturally leads to the question: how do learners perceive TTS in instructional contexts? Since TTS allows controls on speaker characteristics (e.g. gender, regional variety, speed, etc.) the variety of materials that could be created--especially in contexts in which native speakers are difficult or expensive to find--makes this an attractive option. However, the effectiveness of TTS, namely, intelligibility, expressiveness, and naturalness, might be questioned for those instances in which the listening is more empathic than informational. In this study, we examined participants' comprehension of the factual details and speaker emotion as well as collected their opinions towards TTS systems for language learning. This study took place in an intensive English Program (IEP) with an academic focus at a large university in the United States. The participants had ACTFL proficiency levels ranging from Novice High to Advance Low. The participants were divided into two groups and through a counterbalanced design, were given a listening assessment in which half of the listening passages were recorded by voice actors, and other half were generated by the TTS system. After the assessment, the participants were given a survey that inquired their opinion towards TTS systems as learning tools. We did not find significant relationships between the voice delivery and participants' comprehension of details and speakers' emotions. Furthermore, more than half of the participants held positive views to using TTS systems as learning tools; thus, this study suggested the use of TTS systems when applicable.

Det är språkutvecklande att tänka högt, alltså finns det en förbindelse mellan språket och tanken. Studien har utgångspunkter genom kvalitativa intervjuer, lektioner, formativ bedömning, samt själv- och kamratbedömning. Studiens är baserad på den sociokulturella teorin och konstruktivismen. Denna studie syftar till att synliggöra på vilket sätt de olika metoderna kan stödja elevernas språk- och kunskapsutveckling genom berättande text. I enlighet med detta syftar studien att besvara dessa två frågeställningar: Hur kan djupare begreppskunskaper kopplat till berättande text gynna elevernas skrivande? Hur kan skriftlig och muntlig formativ bedömning i undervisningen hjälpa eleverna att utveckla sina berättande texter? För att besvara dessa frågeställningar som formulerats utifrån syftet användes semistrukturerade intervjuer, formativ bedömning under lektionerna, begreppsutveckling från lektionerna, skriftlig formativ bedömning efter elevernas berättande texter, själv- och kamratbedömning som datainsamlingsmetod. Sju stycken slumpmässigt utvalda elever intervjuades. Studiens resultat har påvisat att elevernas föreställningar stämde överens med vad tidigare forskning kommit fram till, gällande formativ bedömning och begreppskunskap. Detta syntes genom elevernas utveckling mellan deras olika berättande texterna som de konstruerade. Sammanfattningsvis visar denna studie att undervisning som fokuserar specifika begrepp i kombination med formativ bedömning kan erbjuda goda möjligheter för språk- och kunskapsutveckling hos eleverna och goda möjligheter för eleverna att nå lärandemålen.

Due to an increase in newer forms of literature’s popularity among the general consumers, research in favor of utilizing media formats such as audiobooks and graphic novels for educational purposes has grown. Literature formats other than the classic printed text have become more accessible in book stores and school libraries during the last decade, making them potentially useful tools for teaching literature in the classrooms. This essay presents several theoretical perspectives on the use of different modalities when working with literature in English teaching at intermediate level, as well as a field study researching the level of engagement between upper secondary school students and storyworlds when working with the same literary work through three different formats: the classic print, the graphic novel and the audiobook. The results, which were obtained through surveys answered by students and interviews with teachers, showed high levels of storyworld engagement from the students in regard to visual media, both when working with the print and the graphic novel, and a significantly lower interest towards the auditory media, i.e. the audiobook.

English title: Another Point of View The aim of this study is to find traces of learning in the process of peer response. This is explored by investigating the experiences expressed by the students and their work of trans­forming criteria into peer response questions. An analysis is also made of which criteria they focus on in their feedback. A close-up study of the meeting between the student, the response group and the text has been made consisting of three students and their groups. The teacher has ranked the groups from her expectations of their performances. Analytical tools are the criteria themselves, metalanguage (Hansson 2001) and local and global text levels (Hoel 2001). The theoretical framework is taken from Ewa Bergh Nestlog and her thesis from 2012. Based on systematic functional linguistics, critical discourse analysis and a dialogical conception of language she presents a transaction cycle. The transaction cycle explains the connection between metaunderstanding, the production and the reception of texts, and it is used in this study for understanding peer response. The study explores how Swedish students in upper secondary school make meaning in peer response. Eighteen students aged 16–17 have digitally been receiving and providing feedback in groups of three while writing an argumentative text. All the students are following a theoretical curriculum. Sources for the material are various written texts by the students. The data analysed consists of the students’ questions to the criteria, their feedback comments on Google Drive and their self-evaluations. The analysis of the self-evaluations shows that receiving and giving feedback helps the students to use a metaperspective while writing and editing their own texts. The students describe how the teaching activities and materials focusing on structure in text have helped them to focus on structure when giving a response. This is confirmed by the analysis of the feedback which shows that the students in their comments give critical and concrete feedback on structure using an adequate language. The results indicate that the chosen focus of teaching activities influences the focus of the students’ response and writing. The results also indicate a need for more teaching activities concerning grammatical terms. The close-up study shows that the students in the groups highly valued by their teacher have an ability to adjust their feedback according to both the text and their peers. The group less valued by their teacher gives very little response and the students do not participate in the self-evaluation.