Disclaimer: This paper contains content that can be perceived as offensive or upsetting. Considerable progress has been made in Artificial intelligence (AI) and Natural language processing (NLP) in the last years. Neural language models (LM) like Generative pre-trained transformer 3 (GPT-3) show impressive results, generating high-quality text seemingly written by a human. Neural language models are already applied in society for example in creating chatbots or assisting with writing documents. As generative LMs are trained on large amounts of data from all kinds of sources, they can pick up toxic traits. GPT-3 has for instance been shown to generate text with social biases, racism, sexism and toxic language. Therefore, filtering for toxic content is necessary to safely deploy models like GPT-3. GPT-3 is trained on and can generate English text data, but similar models for smaller languages have recently emerged. GPT-SWE is a novel model based on the same technical principles as GPT-3, able to generate Swedish text. Much like GPT-3, GPT-SWE has issues with generating toxic text. A promising approach for addressing this problem is to train a separate toxicity classification model for classifying the generated text as either toxic or safe. However, there is a substantial need for more research on toxicity classification for lower resource languages and previous studies for the Swedish language are sparse. This study explores the use of toxicity classifiers to filter Swedish text generated from GPT-SWE. This is investigated by creating and annotating a small Swedish toxicity dataset which is used to fine-tune a Swedish BERT model. The best performing toxicity classifier created in this work cannot be considered useful in an applied scenario. Nevertheless, the results encourage continued studies on BERT models that are pre-trained and fine-tuned in Swedish to create toxicity classifiers. The results also highlight the importance of qualitative datasets for fine-tuning and demonstrate the difficulties of toxicity annotation. Furthermore, expert annotators, distinctive well-defined guidelines for annotation and fine-grained labels are recommended. The study also provides insights into the potential for active learning methods in creating datasets in languages with lower resources. Implications and potential solutions regarding toxicity in generative LMs are also discussed. / Varning: Denna studie omfattar innehåll som kan uppfattas som stötande eller upprörande. Betydande framsteg har gjorts inom Artificiell intelligens (AI) och Språkteknologi (NLP) de senaste åren. Utvecklingen av Neurala språkmodeller har fört med sig framgångsrika modeller likt Generative pre-trained transformer 3 (GPT-3) som visat på imponerande resultat i att generera högkvalitativ text, till synes skriven av en människa. Språkmodeller tillämpas redan på flera platser i samhället till exempel för att hjälpa till med att skriva dokument eller för att skapa chatbots. Eftersom språkmodeller tränas på stora mängder data från många typer av källor kan de fånga upp toxiska egenskaper. GPT-3 har till exempel visat sig generera text med sociala fördomar, rasism, sexism och toxiskt språk. En nödvändighet för att säkert distribuera modeller som GPT-3 inkluderar således filtrering av toxiskt innehåll. GPT-3 är tränad på och kan generera engelsk textdata men liknande modeller för mindre språk har nyligen börjat dyka upp. GPT-SWE är en ny modell som bygger på samma tekniska principer som GPT-3 men kan generera svensk text. Likt GPT-3 så har GPT-SWE problem med genererad toxisk text. För att lösa problemen med toxicitet är ett lovande tillvägagångssätt att träna en separat toxicitetsklassificeringsmodell för att klassificera genererad text som toxisk eller säker. Det finns dock en brist på tidigare studier om detta för det svenska språket och det finns ett stort behov av mer forskning kring toxicitetsklassificering för språk med lägre resurser. Följaktligen undersöker detta projekt möjligheterna att använda toxicitetsklassificerare för att filtrera genererad text från svenska språkmodeller. Detta undersöks genom att skapa och annotera ett litet svenskt toxicitets-dataset som används för att finjustera en svensk BERT-modell. Den bäst presterande toxicitetsklassificeraren som skapades inom detta arbete kan inte anses användbar i ett tillämpat scenario. Resultaten uppmuntrar dock fortsatta studier på BERT-modeller förtränade och finjusterade på svenska för att skapa toxicitetsklassificerare. Resultatet skiftar också ytterligare fokus mot vikten av ett kvalitativt dataset för finjustering och påvisar svårigheterna med toxicitets-annotering. Vidare rekommenderas expert-annoterare, distinkta väldefinierade riktlinjer för annotering samt användandet av fler och mer specificerade kategorier för toxicitet. Arbetet ger dessutom insikter om potentialen för metoder som aktiv inlärning för att skapa dataset inom språk med lägre resurser. Fortsättningsvis diskuteras också implikationer och potentiella lösningar angående toxicitet i språkmodeller.

<p>While instance-dependent contextual bandits have been previously studied, their analysis<br> has been exclusively limited to pure disagreement-based learning. This approach lacks a<br> nuanced understanding of disagreement and treats it in a binary and absolute manner.<br> In our work, we aim to broaden the analysis of instance-dependent contextual bandits by<br> studying them under the framework of disagreement-based learning in sub-regions. This<br> framework allows for a more comprehensive examination of disagreement by considering its<br> varying degrees across different sub-regions.<br> To lay the foundation for our analysis, we introduce key ideas and measures widely<br> studied in the contextual bandit and disagreement-based active learning literature. We<br> then propose a novel, instance-dependent contextual bandit algorithm for the realizable<br> case in a transductive setting. Leveraging the ability to observe contexts in advance, our<br> algorithm employs a sophisticated Linear Programming subroutine to identify and exploit<br> sub-regions effectively. Next, we provide a series of results tying previously introduced<br> complexity measures and offer some insightful discussion on them. Finally, we enhance the<br> existing regret bounds for contextual bandits by integrating the sub-region disagreement<br> coefficient, thereby showcasing significant improvement in performance against the pure<br> disagreement-based approach.<br> In the concluding section of this thesis, we do a brief recap of the work done and suggest<br> potential future directions for further improving contextual bandit algorithms within the<br> framework of disagreement-based learning in sub-regions. These directions offer opportuni-<br> ties for further research and development, aiming to refine and enhance the effectiveness of<br> contextual bandit algorithms in practical applications.<br> <br> </p>

In this work, we set the stage of a limited labelling budget and propose using a sample selector network to learn and select effective training samples, whose labels we would then acquire to train the target model performing the required machine learning task. We make the assumption that the sample features, the state of the target model and the training loss of the target model are informative for training the sample selector network. In addition, we approximate the state of the target model with its intermediate and final network outputs. We investigate if under a limited labelling budget, the sample selector network is capable of learning and selecting training samples that train the target model at least as effectively as using another training subset of the same size that is uniformly randomly sampled from the full training dataset, the latter being the common procedure used to train machine learning models without active learning. We refer to this common procedure as the traditional machine learning uniform random sampling method. We perform experiments on the MNIST and CIFAR-10 datasets; and demonstrate with empirical evidence that under a constrained labelling budget and some other conditions, active learning using a sample selector network enables the target model to learn more effectively. / I detta arbete sätter vi steget i en begränsad märkningsbudget och föreslår att vi använder ett provväljarnätverk för att lära och välja effektiva träningsprover, vars etiketter vi sedan skulle skaffa för att träna målmodellen som utför den nödvändiga maskininlärningsuppgiften. Vi antar att provfunktionerna, tillståndet för målmodellen och utbildningsförlusten för målmodellen är informativa för att träna provväljarnätverket. Dessutom uppskattar vi målmodellens tillstånd med dess mellanliggande och slutliga nätverksutgångar. Vi undersöker om provväljarnätverket enligt en begränsad märkningsbudget kan lära sig och välja utbildningsprover som tränar målmodellen minst lika effektivt som att använda en annan träningsdel av samma storlek som är enhetligt slumpmässigt samplad från hela utbildningsdatasystemet, det senare är det vanliga förfarandet som används för att utbilda maskininlärningsmodeller utan aktivt lärande. Vi hänvisar till denna vanliga procedur som den traditionella maskininlärning enhetliga slumpmässig sampling metod. Vi utför experiment på datasätten MNIST och CIFAR-10; och visa med empiriska bevis att under en begränsad märkningsbudget och vissa andra förhållanden, aktivt lärande med hjälp av ett provvalnätverk gör det möjligt för målmodellen att lära sig mer effektivt.

The recent advancements of Natural Language Processing have cleared the path for many new applications. This is primarily a consequence of the transformer model and the transfer-learning capabilities provided by models like BERT. However, task-specific labeled data is required to fine-tune these models. To alleviate the expensive process of labeling data, Active Learning (AL) aims to maximize the information gained from each label. By including a model in the annotation process, the informativeness of each unlabeled sample can be estimated and hence allow human annotators to focus on vital samples and avoid redundancy. This thesis investigates to what extent AL can accelerate model training with respect to the number of labels required. In particular, the focus is on pre- trained Swedish language models in the context of Named Entity Recognition. The data annotation process is simulated using existing labeled datasets to evaluate multiple AL strategies. Experiments are evaluated by analyzing the F1 score achieved by models trained on the data selected by each strategy. The results show that AL can significantly accelerate the model training and hence reduce the manual annotation effort. The state-of-the-art strategy for sentence classification, ALPS, shows no sign of accelerating the model training. However, uncertainty-based strategies consistently outperform random selection. Under certain conditions, these strategies can reduce the number of labels required by more than a factor of two. / Framstegen som nyligen har gjorts inom naturlig språkbehandling har möjliggjort många nya applikationer. Det är mestadels till följd av transformer-modellerna och lärandeöverföringsmöjligheterna som kommer med modeller som BERT. Däremot behövs det fortfarande uppgiftsspecifik annoterad data för att finjustera dessa modeller. För att lindra den dyra processen att annotera data, strävar aktiv inlärning efter att maximera informationen som utvinns i varje annotering. Genom att inkludera modellen i annoteringsprocessen, kan man estimera hur informationsrikt varje träningsexempel är, och på så sätt låta mänskilga annoterare fokusera på viktiga datapunkter. Detta examensarbete utforskar hur väl aktiv inlärning kan accelerera modellträningen med avseende på hur många annoterade träningsexempel som behövs. Fokus ligger på förtränade svenska språkmodeller och uppgiften namnigenkänning. Dataannoteringsprocessen simuleras med färdigannoterade dataset för att evaluera flera olika strategier för aktiv inlärning. Experimenten evalueras genom att analysera den uppnådda F1-poängen av modeller som är tränade på datapunkterna som varje strategi har valt. Resultaten visar att aktiv inlärning har en signifikant förmåga att accelerera modellträningen och reducera de manuella annoteringskostnaderna. Den toppmoderna strategin för meningsklassificering, ALPS, visar inget tecken på att kunna accelerera modellträningen. Däremot är osäkerhetsbaserade strategier är konsekvent bättre än att slumpmässigt välja datapunkter. I vissa förhållanden kan dessa strategier reducera antalet annoteringar med mer än en faktor 2.

Data labelling for question answering tasks (QA) is a costly procedure that requires oracles to read lengthy excerpts of texts and reason to extract an answer for a given question from within the text. QA is a task in natural language processing (NLP), where a majority of recent advancements have come from leveraging the vast corpora of unlabelled and unstructured text available online. This work aims to extend this trend in the efficient use of unlabelled text data to the problem of selecting which subset of samples to label in order to maximize performance. This practice of selective labelling is called active learning (AL).  Recent developments in AL for NLP have introduced the use of self-supervised learning on large corpora of text in the labelling process of samples for classification problems. This work adapts this research to the task of question answering and performs an initial exploration of expected performance.  The methods covered in this work use uncertainty estimates obtained from neural networks to guide an incremental labelling process. These estimates are obtained from transformer-based models, previously trained in a self-supervised manner, by calculating the entropy of the confidence scores or with an approximation of Bayesian uncertainty obtained through Monte Carlo dropout. These methods are evaluated on two different benchmarking QA datasets: SQuAD v1 and TriviaQA.  Several factors are observed to influence the behaviour of these uncertainty-based acquisition functions, including the choice of language model used, the presence of unanswered questions and the acquisition size used in the incremental process. The study produces no evidence to support that averaging or selecting maximal uncertainty values between the classification of an answer’s starting and ending positions affects sample acquisition quality. However, language model choice, the presence of unanswerable questions and acquisition size are all identified as key factors affecting consistency between runs and degree of success.

Die modellbasierte Kalibrierung dynamischer Fahrmanöver an Prüfständen ermöglicht die systematische Optimierung von Steuergerätedaten über den gesamten Betriebsbereich des Fahrzeugs und begegnet somit der steigenden Komplexität in der Antriebsstrangentwicklung. Dabei werden mehrere empirische Black-Box-Modelle zur Abbildung der Zielgrößen für die nachfolgende Optimierung identifiziert. Der Einsatz der statistischen Versuchsplanung ermöglicht eine systematische Abdeckung des gesamten Eingangsbereiches. In jüngerer Vergangenheit werden in der Automobilindustrie vereinzelt Methoden des maschinellen Lernens eingesetzt, um die Anwendung der modellbasierten Kalibrierung zu vereinfachen und die Effizienz zu erhöhen. Insbesondere der Einsatz des aktiven Lernens führt zu vielversprechenden Ergebnissen. Mit diesen Methoden werden Modelle mit einer geringeren Anzahl an Messpunkten identifiziert, während gleichzeitig die erforderliche Expertise für die Versuchsplanerstellung reduziert wird. Eine Herausforderung stellt die simultane Identifikation mehrerer Regressionsmodelle dar, die für die Anwendung des aktiven Lernens auf die Fahrbarkeitskalibrierung erforderlich ist. Hierfür wird im Rahmen dieser Arbeit die aktive Ausgangsselektion (AOS) eingeführt und eingesetzt. Die AOS-Strategie bestimmt dabei das führende Modell im Lernprozess. Erste Veröffentlichungen zeigen das Potenzial der Verwendung von AOS. Statistisch signifikante Ergebnisse über die Effektivität gibt es bislang jedoch nicht, weswegen die weitere intensive Untersuchung von Strategien erforderlich ist. In der vorliegenden Arbeit werden regel- und informationsbasierte AOS-Strategien vorgestellt. Letztere wählen das führende Modell basierend auf allen während des Versuchs verfügbaren Informationen aus. Hier erfolgt erstmals die detaillierte Beschreibung und Untersuchung einer normierten modellgütebasierten Auswahlstrategie. Als Modellart werden Gauß’sche Prozessmodelle verwendet. Anhand von Versuchen wird überprüft, ob der Einsatz von AOS gegenüber gängiger statistischer Versuchsplanung sinnvoll ist. Darüber hinaus wird untersucht, ob die Berücksichtigung aller zur Versuchslaufzeit bekannten Informationen zu einer Verbesserung des Lernprozesses beiträgt. Die Strategien werden an Simulationsexperimenten getestet. Diese Simulationsexperimente stellen Grenzfälle echter Versuche dar, die für die Strategien besonders herausfordernd sind. Die Erstellung der Experimente wird anhand von Informationen aus realen Prüfstandsversuchen abgeleitet. Die Strategien werden analysiert und miteinander verglichen. Dazu wird eine anspruchsvolle Referenzstrategie verwendet, die auf den Methoden der klassischen Versuchsplanung basiert. Die Versuche zeigen, dass bereits einfache regelbasierte Strategien bessere Ergebnisse hervorbringen als die Referenzstrategie. Durch Berücksichtigung der momentanen Modellgüte und Abschätzung des Prozessrauschens zur Versuchslaufzeit ist eine weitere Reduktion der Messpunkte um mehr als 50% gegenüber der Referenzstrategie möglich. Da die informationsbasierte Strategie rechenintensiver ist, wird auch ein zeitlicher Vergleich mit unterschiedlichen langen Annahmen für die Fahrmanöverdauer am Prüfstand vorgenommen. Bei kurzen Manöverzeiten ist der Vorteil der informationsbasierten Strategie gegenüber der regelbasierten Strategie nur gering ausgeprägt. Mit zunehmender Manöverzeit nähert sich die abgeschätzte zeitliche Ersparnis jedoch der prozentualen Einsparung der Messpunkte an. Die aus den Simulationsexperimenten abgeleiteten Ergebnisse werden anhand eines realen Anwendungsbeispiels validiert. Die Implementierung an einem Antriebsstrangprüfstand wird dazu vorgestellt. Für die Versuche werden insgesamt 1500 Fahrmanöver an diesem Prüfstand durchgeführt. Die Ergebnisse der Versuche bestätigen die aus den Simulationsexperimenten abgeleiteten Ergebnisse. Die regelbasierte AOS-Strategie reduziert die Anzahl der Messpunkte im Durchschnitt um 65% im Vergleich zur verwendeten Referenzstrategie. Die informationsbasierte AOS-Strategie verringert die Anzahl der Punkte weiter auf 70% gegenüber der Referenzstrategie. Die Modelle der informationsbasierten Strategie sind bereits nach 50% der Punkte besser als die besten Modelle der regelbasierten Strategie. Die Ergebnisse dieser Arbeit legen den ständigen Einsatz der vorgestellten informationsbasierten Strategien für die modellbasierte Kalibrierung nahe. / Model-based calibration of dynamic driving maneuvers on test benches enables the systematic optimization of ECU data over the vehicle’s entire operating range and thus faces the increasing complexity in powertrain development. Several empirical black-box models are identified to represent the target variables for the succeeding optimization. The use of statistical experimental design enables systematic coverage of the entire input range. Recently, machine learning methods have been occasionally used in the automotive industry to simplify applying the process and increase its efficiency. In particular, the use of active learning leads to promising results. It leads to a reduction of the number of measurement points necessary for model identification. At the same time, the required expertise for experimental design is reduced. The simultaneous identification of multiple regression models, which is required for a broad application of active learning to drivability calibration, is challenging. In this work, active output selection (AOS) is introduced and applied to face this challenge. An AOS strategy determines the leading model in the learning process. First publications show the potential of using AOS. However, no statistically significant results about the effectiveness are available to date, which is why these strategies need to be studied in more detail. This work presents rule- and information-based AOS strategies. The latter select the leading model based on all current information available during the experiment. For the first time, this publication provides a detailed description and investigation of a normalized model-quality-based selection strategy. Gaussian process models are used as model type. Experiments are conducted to verify whether the use of AOS is reasonable compared to common designs of experiments. Furthermore, we analyze whether taking into account all information known at the time of the experiment helps to improve the learning process. The strategies are first tested on computer experiments. These computer experiments represent borderline cases of real experiments, which are particularly challenging for the strategies. The experiments are derived using information from real test bench experiments. The strategies are analyzed and compared with each other. For this purpose, a sophisticated reference strategy is used, which is based on the methods of classical designs of experiments. The experiments show that even simple rule-based strategies lead to better results than the reference strategy. By considering the current model quality and estimating the process noise during experiment runtime, a further reduction of the measurement points by more than 50% compared to the reference strategy is possible. Since the information-based strategy is more computationally expensive, we perform a time comparison with different assumptions for the driving maneuver duration at the test bench. For short maneuver times, the advantage of the information-based strategy in comparison to the rule-based strategy is only small. As the maneuver time increases, the estimated time reduction approaches the percentage savings of the measurement points. The results derived from the computer experiments are validated using a real application example. The implementation on a powertrain test bench is presented for this purpose. For the experiments, a total of 1500 driving maneuvers are performed on this test bench. The results of the experiments confirm the results of the computer experiments. The rule-based AOS strategy reduces the number of measurement points by 65% on average compared to the reference strategy used. The information-based AOS strategy further reduces the number of points to 70% compared to the reference strategy. The results of this work suggest the use of the presented information-based strategies for model-based calibration.

Additive manufacturing (AM) technology is a key emerging field transforming how customized products with complex shapes are manufactured. AM is the process of layering materials to produce objects from three-dimensional (3D) models. AM technology can be used to print objects with complicated geometries and a broad range of material properties. However, the issue of ensuring the quality of printed products during the process remains an obstacle to industry-level adoption. Furthermore, the characteristics of AM processes typically involve complex process dynamics and interactions between machine parameters and desired qualities. The issues associated with quality assurance in AM processes underscore the need for research into smart quality assurance systems. To study the complex physics behind process interaction challenges in AM processes, this dissertation proposes the development of a data-driven smart quality assurance framework that incorporates in-process sensing and machine learning-based modeling by correlating the relationships among parameters, signatures, and quality. High-fidelity AM simulation data and the increasing use of sensors in AM processes help simulate and monitor the occurrence of defects during a process and open doors for data-driven approaches such as machine learning to make inferences about quality and predict possible failure consequences. To address the research gaps associated with quality assurance for AM processes, this dissertation proposes several data-driven approaches based on the design of experiments (DoE), forward prediction modeling, and an inverse design methodology. The proposed approaches were validated for AM processes such as fused filament fabrication (FFF) using polymer and hydrogel materials and laser powder bed fusion (LPBF) using common metal materials. The following three novel smart quality assurance systems based on a parameter–signature–quality (PSQ) framework are proposed: 1. A customized in-process sensing platform with a DOE-based process optimization approach was proposed to learn and optimize the relationships among process parameters, process signatures, and parts quality during bioprinting processes. This approach was applied to layer porosity quantification and quality assurance for polymer and hydrogel scaffold printing using an FFF process. 2. A data-driven surrogate model that can be informed using high-fidelity physical-based modeling was proposed to develop a parameter–signature–quality framework for the forward prediction problem of estimating the quality of metal additive-printed parts. The framework was applied to residual stress prediction for metal parts based on process parameters and thermal history with reheating effects simulated for the LPBF process. 3. Deep-ensemble-based neural networks with active learning for predicting and recommending a set of optimal process parameter values were developed to optimize optimal process parameter values for achieving the inverse design of desired mechanical responses of final built parts in metal AM processes with fewer training samples. The methodology was applied to metal AM process simulation in which the optimal process parameter values of multiple desired mechanical responses are recommended based on a smaller number of simulation samples. / Doctor of Philosophy / Additive manufacturing (AM) is the process of layering materials to produce objects from three-dimensional (3D) models. AM technology can be used to print objects with complicated geometries and a broad range of material properties. However, the issue of ensuring the quality of printed products during the process remains a challenge to industry-level adoption. Furthermore, the characteristics of AM processes typically involve complex process dynamics and interactions between machine parameters and the desired quality. The issues associated with quality assurance in AM processes underscore the need for research into smart quality assurance systems. To study the complex physics behind process interaction challenges in AM processes, this dissertation proposes a data-driven smart quality assurance framework that incorporates in-process sensing and machine-learning-based modeling by correlating the relationships among process parameters, sensor signatures, and parts quality. Several data-driven approaches based on the design of experiments (DoE), forward prediction modeling, and an inverse design methodology are proposed to address the research gaps associated with implementing a smart quality assurance system for AM processes. The proposed parameter–signature–quality (PSQ) framework was validated using bioprinting and metal AM processes for printing with polymer, hydrogel, and metal materials.

Educational Digital Libraries (DLs) are complex information systems which are designed to support individuals' information needs and information seeking behavior. To have a broad impact on the communities in education and to serve for a long period, DLs need to structure and organize the resources in a way that facilitates the dissemination and the reuse of resources. Such a digital library should meet defined quality dimensions in the 5S (Societies, Scenarios, Spaces, Structures, Streams) framework - including completeness, consistency, efficiency, extensibility, and reliability - to ensure that a good quality DL is built. In this research, we addressed both external and internal quality aspects of DLs. For internal qualities, we focused on completeness and consistency of the collection, catalog, and repository. We developed an application pipeline to acquire user-generated computing-related resources from YouTube and SlideShare for an educational DL. We applied machine learning techniques to transfer what we learned from the ACM Digital Library dataset. We built classifiers to catalog resources according to the ACM Computing Classification System from the two new domains that were evaluated using Amazon Mechanical Turk. For external qualities, we focused on efficiency, scalability, and reliability in DL services. We proposed cloud-based designs and applications to ensure and improve these qualities in DL services using cloud computing. The experimental results show that our proposed methods are promising for enhancing and enriching an educational digital library. This work received support from ACM, as well as the National Science Foundation under Grant Numbers DUE-0836940, DUE-0937863, and DUE-0840719, and IMLS LG-71-16-0037-16. / Ph. D.

We present an active learning algorithm for identifying viable process conditions in magnetron sputtering experiments. The algorithm trains a classifier to predict which gas pressure and magnetron power combinations yield stable discharge with deposition rates exceeding a minimum threshold. A computation-based oracle that labels experiments using QCM readings facilitates a fully automated learning procedure, laying the groundwork for a self-driving lab where novel materials will be explored for next-generation solar cells. Upon evaluation, the active learning algorithm results in significantly higher sample efficiency than traditional supervised learning across a range of magnetron sputtering experiments. Moreover, a sampling sequence analysis shows that active learning enables an informed search of the process parameter space, generating patterns that approximate Paschen’s law. The work presented in this thesis serves as a first step toward a fully automated materials synthesis process, where the input region of viable synthesis parameters can be identified with minimal experimentation. The solution allows researchers to efficiently narrow the search space of optimal process conditions for targeted materials design.

This dissertation comprises several studies addressing supervised learning problems where the supervision is imperfect. Firstly, we investigate the margin conditions in active learning. Active learning is characterized by its special mechanism where the learner can sample freely over the feature space and exploit mostly the limited labeling budget by querying the most informative labels. Our primary focus is to discern critical conditions under which certain active learning algorithms can outperform the optimal passive learning minimax rate. Within a non-parametric multi-class classification framework,our results reveal that the uniqueness of Bayes labels across the feature space serves as the pivotal determinant for the superiority of active learning over passive learning. Secondly, we study the estimation of central mean subspace (CMS), and its application in transfer learning. We show that a fast parametric convergence rate is achievable via estimating the expected smoothed gradient outer product, for a general class of covariate distribution that admits Gaussian or heavier distributions. When the link function is a polynomial with a degree of at most r and the covariates follow the standard Gaussian, we show that the prefactor depends on the ambient dimension d as d^r. Furthermore, we show that under a transfer learning setting, an oracle rate of prediction error as if the CMS is known is achievable, when the source training data is abundant. Finally, we present an innovative application involving the utilization of weak (noisy) labels for addressing an Individual Tree Crown (ITC) segmentation challenge. Here, the objective is to delineate individual tree crowns within a 3D LiDAR scan of tropical forests, with only 2D noisy manual delineations of crowns on RGB images available as a source of weak supervision. We propose a refinement algorithm designed to enhance the performance of existing unsupervised learning methodologies for the ITC segmentation problem.