Search

Your search keyword '"Vogt, Stephan"' showing total 564 results
Start Over Author "Vogt, Stephan"
564 results on '"Vogt, Stephan"'

4. Targeted Adversarial Attacks on Wind Power Forecasts

Author

Heinrich, René, Scholz, Christoph, Vogt, Stephan, and Lehna, Malte

Subjects
Computer Science - Machine Learning, Computer Science - Cryptography and Security
Abstract

In recent years, researchers proposed a variety of deep learning models for wind power forecasting. These models predict the wind power generation of wind farms or entire regions more accurately than traditional machine learning algorithms or physical models. However, latest research has shown that deep learning models can often be manipulated by adversarial attacks. Since wind power forecasts are essential for the stability of modern power systems, it is important to protect them from this threat. In this work, we investigate the vulnerability of two different forecasting models to targeted, semi-targeted, and untargeted adversarial attacks. We consider a Long Short-Term Memory (LSTM) network for predicting the power generation of individual wind farms and a Convolutional Neural Network (CNN) for forecasting the wind power generation throughout Germany. Moreover, we propose the Total Adversarial Robustness Score (TARS), an evaluation metric for quantifying the robustness of regression models to targeted and semi-targeted adversarial attacks. It assesses the impact of attacks on the model's performance, as well as the extent to which the attacker's goal was achieved, by assigning a score between 0 (very vulnerable) and 1 (very robust). In our experiments, the LSTM forecasting model was fairly robust and achieved a TARS value of over 0.78 for all adversarial attacks investigated. The CNN forecasting model only achieved TARS values below 0.10 when trained ordinarily, and was thus very vulnerable. Yet, its robustness could be significantly improved by adversarial training, which always resulted in a TARS above 0.46., Comment: 21 pages, including appendix, 12 figures

Published
2023

5. Efficient Bayesian Updates for Deep Learning via Laplace Approximations

Author

Huseljic, Denis, Herde, Marek, Rauch, Lukas, Hahn, Paul, Huang, Zhixin, Kottke, Daniel, Vogt, Stephan, and Sick, Bernhard

Subjects
Computer Science - Machine Learning
Abstract

Since training deep neural networks takes significant computational resources, extending the training dataset with new data is difficult, as it typically requires complete retraining. Moreover, specific applications do not allow costly retraining due to time or computational constraints. We address this issue by proposing a novel Bayesian update method for deep neural networks by using a last-layer Laplace approximation. Concretely, we leverage second-order optimization techniques on the Gaussian posterior distribution of a Laplace approximation, computing the inverse Hessian matrix in closed form. This way, our method allows for fast and effective updates upon the arrival of new data in a stationary setting. A large-scale evaluation study across different data modalities confirms that our updates are a fast and competitive alternative to costly retraining. Furthermore, we demonstrate its applicability in a deep active learning scenario by using our update to improve existing selection strategies., Comment: submitted to NeurIPS

Published
2022

7. Task Embedding Temporal Convolution Networks for Transfer Learning Problems in Renewable Power Time-Series Forecast

Author

Schreiber, Jens, Vogt, Stephan, and Sick, Bernhard

Subjects
Computer Science - Machine Learning
Abstract

Task embeddings in multi-layer perceptrons for multi-task learning and inductive transfer learning in renewable power forecasts have recently been introduced. In many cases, this approach improves the forecast error and reduces the required training data. However, it does not take the seasonal influences in power forecasts within a day into account, i.e., the diurnal cycle. Therefore, we extended this idea to temporal convolutional networks to consider those seasonalities. We propose transforming the embedding space, which contains the latent similarities between tasks, through convolution and providing these results to the network's residual block. The proposed architecture significantly improves up to 25 percent for multi-task learning for power forecasts on the EuropeWindFarm and GermanSolarFarm dataset compared to the multi-layer perceptron approach. Based on the same data, we achieve a ten percent improvement for the wind datasets and more than 20 percent in most cases for the solar dataset for inductive transfer learning without catastrophic forgetting. Finally, we are the first proposing zero-shot learning for renewable power forecasts to provide predictions even if no training data is available., Comment: Accepted at European Conference on Machine Learning 2021

Published
2022
Full Text
View/download PDF

8. Synthetic Photovoltaic and Wind Power Forecasting Data

Author

Vogt, Stephan, Schreiber, Jens, and Sick, Bernhard

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Electrical Engineering and Systems Science - Signal Processing
Abstract

Photovoltaic and wind power forecasts in power systems with a high share of renewable energy are essential in several applications. These include stable grid operation, profitable power trading, and forward-looking system planning. However, there is a lack of publicly available datasets for research on machine learning based prediction methods. This paper provides an openly accessible time series dataset with realistic synthetic power data. Other publicly and non-publicly available datasets often lack precise geographic coordinates, timestamps, or static power plant information, e.g., to protect business secrets. On the opposite, this dataset provides these. The dataset comprises 120 photovoltaic and 273 wind power plants with distinct sides all over Germany from 500 days in hourly resolution. This large number of available sides allows forecasting experiments to include spatial correlations and run experiments in transfer and multi-task learning. It includes side-specific, power source-dependent, non-synthetic input features from the ICON-EU weather model. A simulation of virtual power plants with physical models and actual meteorological measurements provides realistic synthetic power measurement time series. These time series correspond to the power output of virtual power plants at the location of the respective weather measurements. Since the synthetic time series are based exclusively on weather measurements, possible errors in the weather forecast are comparable to those in actual power data. In addition to the data description, we evaluate the quality of weather-prediction-based power forecasts by comparing simplified physical models and a machine learning model. This experiment shows that forecasts errors on the synthetic power data are comparable to real-world historical power measurements., Comment: 12 pages, 8 figures, and 2 tables

Published
2022

9. Quantile Surfaces -- Generalizing Quantile Regression to Multivariate Targets

Author

Bieshaar, Maarten, Schreiber, Jens, Vogt, Stephan, Gensler, André, and Sick, Bernhard

Subjects
Statistics - Applications, Computer Science - Artificial Intelligence, Statistics - Computation
Abstract

In this article, we present a novel approach to multivariate probabilistic forecasting. Our approach is based on an extension of single-output quantile regression (QR) to multivariate-targets, called quantile surfaces (QS). QS uses a simple yet compelling idea of indexing observations of a probabilistic forecast through direction and vector length to estimate a central tendency. We extend the single-output QR technique to multivariate probabilistic targets. QS efficiently models dependencies in multivariate target variables and represents probability distributions through discrete quantile levels. Therefore, we present a novel two-stage process. In the first stage, we perform a deterministic point forecast (i.e., central tendency estimation). Subsequently, we model the prediction uncertainty using QS involving neural networks called quantile surface regression neural networks (QSNN). Additionally, we introduce new methods for efficient and straightforward evaluation of the reliability and sharpness of the issued probabilistic QS predictions. We complement this by the directional extension of the Continuous Ranked Probability Score (CRPS) score. Finally, we evaluate our novel approach on synthetic data and two currently researched real-world challenges in two different domains: First, probabilistic forecasting for renewable energy power generation, second, short-term cyclists trajectory forecasting for autonomously driving vehicles. Especially for the latter, our empirical results show that even a simple one-layer QSNN outperforms traditional parametric multivariate forecasting techniques, thus improving the state-of-the-art performance., Comment: Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), currently under review, 15 page, 23 figures, 2 tables

Published
2020

13. Einfluss von Meniskusnähten auf die Knorpelbeschaffenheit von Kniegelenken am Schweinemodell

Author

Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Burgkart, Rainer (Prof. Dr.), Henke, Ferdinand Konstantin, Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Burgkart, Rainer (Prof. Dr.), and Henke, Ferdinand Konstantin

Abstract

In der vorliegenden Arbeit sollte untersucht werden, inwiefern gängige Fadenmaterialien, welche bei der Versorgung von Meniskus-Verletzungen eingesetzt werden, Schäden am Kniegelenksknorpel hinterlassen können. Hierfür wurden drei verschiedene Fäden an Kniegelenken frisch geschlachteter Schweine biomechanisch getestet. Im Anschluss wurden die entstandenen Schäden auf den korrespondierenden Knorpelflächen makroskopisch (Tuschfärbung) und mikroskopisch (Histologie) ausgewertet., The present dissertation should investigate if commonly used suture materials in meniscus injury repair may cause damage to knee joint cartilage. Three different sutures were biomechanically tested on the knee joints of freshly slaughtered pigs. Subsequently, the resulting damage on the corresponding cartilage surfaces was evaluated macroscopically (Indian ink staining) and microscopically (histology).

Published
2024

14. Iliac Crest Bone Block Autograft Transfer for Ballistic Posterior Glenoid Fracture: A Case Report.

Author

Serotte, Jordan Cook, Baker, Hayden, Lee, Cody, Strelzow, Jason A., and Vogt, Stephan

Abstract

Case: A 25‐year‐old male presented with a ballistic fracture of the right glenoid resulting in > 30% loss of the posterior glenoid articular surface and acute posterior glenohumeral instability that was treated with open reduction internal fixation with iliac crest autograft transfer. Conclusion: There is limited consensus on the operative management of ballistic intra‐articular fractures due to the heterogeneity of these injuries. Acute posterior glenohumeral instability secondary to a ballistic fracture is a rare injury pattern. In this case, we were able to successfully treat posterior glenohumeral instability with iliac crest autograft transfer and open reduction internal fixation. [ABSTRACT FROM AUTHOR]

Published
2025
Full Text
View/download PDF

15. The lack of retropatellar resurfacing at index surgery is significantly associated with failure in patients following patellofemoral inlay arthroplasty: a multi-center study of more than 260 patients

Author

Imhoff, Andreas B., Bartsch, Eva, Becher, Christoph, Behrens, Peter, Bode, Gerrit, Cotic, Matthias, Diermeier, Theresa, Falk, Holger, Feucht, Matthias J., Haupt, Ulrich, Hinterwimmer, Stefan, Holz, Johannes, Hutter, René, Kaiser, René, Knoblauch, Tobias, Nebelung, Wolfgang, Niemeyer, Philipp, O’Donnel, Turlough, Pagenstert, Geert, Patzer, Thilo, Rose, Tim, Rupp, Marco C., Tischer, Thomas, Venjakob, Arne J., Vogt, Stephan, and Pogorzelski, Jonas

Published
2022
Full Text
View/download PDF

24. Spontanheilungsrate und klinische Stabilität nach konservativ versorgter Ruptur des vorderen Kreuzbandes

Author

Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);von Eisenhart-Rothe, Rüdiger (Prof. Dr.), Trinnes, Klaus, Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);von Eisenhart-Rothe, Rüdiger (Prof. Dr.), and Trinnes, Klaus

Abstract

Rupturen des VKB werden in aller Regel operativ rekonstruiert. Trotzdem zeigen einige Patienten normale Kniefunktion ohne die Rekonstruktion des VKB.Ziel der Studie war das Aufzeigen von Prognosefaktoren für eine Spontanheilung sowie die Messung des funktionellen Outcomes bei entsprechenden Patienten.Spontanheilungen zeigten sich v.a. bei Patienten niedrigen sportlichen Anspruches mit femoralen Bandabrissen und nicht erhöhtem tibialen Slope. In diesen Fällen sollte ein abwartendes Procedere erwogen werden um die Chance der Spontanheilung zu gewähren., Anterior cruciate ligament (ACL) reconstruction is considered the first line treatment in ACL rupture.However, some patients show normal knee function without ACL reconstruction. Aim of this study was to evaluate the rate and prognostic factors of spontaneous healing as well as the functional outcome in patients with ACL rupture.Spontaneous healing of a ruptured ACL happened especially in low-demand patients with femoral ruptures without increased posterior tibial slope. Delayed ACL surgery should be considered to await the possibility for potential spontaneous ACL healing.

Published
2023

32. Der Einfluss von anatomischen Risikofaktoren auf das Behandlungsergebnis nach Rekonstruktion des medialen patellofamoralen Ligaments

Author

Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Siebenlist, Sebastian (Prof. Dr.), Watermann, Kathrin, Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Siebenlist, Sebastian (Prof. Dr.), and Watermann, Kathrin

Abstract

Diese Studie soll klären, ob die alleinige Behandlung von ligamentären Einflussfaktoren (MPFL) bei chronischer patellofemoraler Instabilität ausreicht, um gute klinische Ergebnisse zu erzielen, oder ob andere Risikofaktoren die Ergebnisse negativ beeinflussen. Als Ergebnis zeigte sich, dass die isolierte MPFL-Rekonstruktion bei Patienten mit patellofemoraler Instabilität und leichten bis mittelschweren Risikofaktoren ein wirksames Verfahren sein kann., This study is intended to clarify whether addressing only ligamentous influence factors (MPFL) in chronic patellofemoral instability is sufficient to produce good clinical outcomes, or whether other risk factors influence the results negatively. As a result, isolated MPFL reconstruction was shown to be an effective procedure in patients with patellofemoral instability and mild to moderate risk factors.

Published
2022

33. Kontrastmittelgestützte, zelluläre Magnetresonanztomographie zum in vivo Monitoring markierter mesenchymaler Stammzellen nach Implantation in osteochondrale Defekte im Kaninchenmodell

Author

Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Makowski, Marcus (Prof. Dr.), Masius, Stephan, Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Makowski, Marcus (Prof. Dr.), and Masius, Stephan

Abstract

Knorpelschäden spielen eine zunehmende Rolle in den Industrienationen. Diese Arbeit untersuchte die zelluläre MRT als Möglichkeit des in vivo Monitorings implantierter MSC im Hinblick auf eine Langzeit-Beobachtung. Hierzu wurden allogene MSC, welche zur Hälfte mit Resovist markiert wurden, in osteochondrale Defekte im Kaninchenmodell implantiert. In folgenden MRT Untersuchungen wurde ein signifikanter Unterschied des SNR, abhängig von der gewählten Sequenz, bis zu 24 Wochen gezeigt., Cartilage defects have an increasing role in industrial nations. This study investigated cellular MRI as a method for in-vivo monitoring of implanted MSC for long-term monitoring. Therefor allogenic MSC, which were one half each labelled with Resovist, were implanted in osteochondral defects in a rabbit model. There was a significant difference in the SNR in following MRI scans, depending on the sequence up to 24 weeks.

Published
2022

34. Klinische und MR-tomographische Ergebnisse nach arthroskopischer Refixation von ossären Bankart-Läsionen mit Fadenankern

Author

Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Makowski, Marcus (Prof. Dr.), Bangoj, Robert, Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Makowski, Marcus (Prof. Dr.), and Bangoj, Robert

Abstract

In der vorliegenden Arbeit konnten 45 Patienten nachuntersucht werden, welche sich zwischen November 1999 und April 2010 in der Abteilung für Sportorthopädie des Klinikums rechts der Isar einer arthroskopischen ossären Bankart-Fadenanker-Refixation unterzogen haben. Zielsetzung dieser Arbeit war es, die mittelfristigen Ergebnisse nach arthroskopischer Refixation von ossären Bankart-Läsionen anhand von einem eigens erstellten Fragebogen, anhand standardisierter und validierter Scores sowie MRT-Aufnahmen zu beschreiben. Weiterhin sollte ein möglicher Zusammenhang zwischen dem MR-tomographischen und dem klinischen Ergebnis untersucht werden. Das Patientenkollektiv bestand aus 20 % weiblichen und 80 % männlichen Patienten. Als Mechanismus für die Erstluxationen gaben 80 % der Patienten einen Fall auf den ausgestreckten Arm an. Im Vergleich zu den zuvor publizierten Studien, lag das Patientenalter in unserer Studie zum Zeitpunkt der operativen Versorgung im Durchschnitt bei 41,2 ± 15,1 Jahren. Unser ,,Follow-up‘‘ betrug im Mittel 82 Monate. Die erhobenen Daten zeigen, dass mit der beschriebenen Technik gute mittelfristige Ergebnisse bezüglich der klinischen Tests, der Rezidiv-Rate und der Patientenzufriedenheit erreicht werden. Die Rekonstruktion des Glenoids konnte in den meisten Fällen anatomisch oder in nur minimaler Fehlstellung des Fragments und ohne relevanten Glenoiddefekt erreicht werden. Die Rezidiv-Rate bei unserem Patientenkollektiv lag bei 6,6 %, wobei 2 von den 3 Patienten eine traumatische Reluxation als Rezidiv-Grund angaben. Ein Anteil von 95 % unserer Patienten konnte postoperativ erneut ihren sportlichen Aktivitäten nachgehen. Jedoch zeigte sich eine signifikante Reduktion des Sportumfangs für alle erhobenen Werte. Als Grund für die Reduktion des Aktivitäts- und Risikolevels wurden zumeist nicht schulter-spezifische Gründe angegeben. So gab fast die Hälfte (44 %) der von uns nachuntersuchten Patienten an, trotz stabiler und schmerzfreier Schulter Sorgen, The purpose of this study was to evaluate the outcome, return to sporting activity, and postoperative articular cartilage and bony morphology of shoulders that underwent arthroscopic suture anchor repair of bony Bankart lesions. The inclusion criteria for this retrospective study were anterior glenoid rim fractures after traumatic shoulder instability that were treated with arthroscopic suture anchor repair. Patients were surveyed by a questionnaire including sport-specific outcome, Rowe score, Western Ontario Shoulder Instability Index, and Oxford Instability Score. Three- tesla magnetic resonance imaging could be performed in 30 patients to assess osseous integration, glenoid reconstruc- tion, and signs of osteoarthritis. Results: From November 1999 to April 2010, 81 patients underwent an anterior bony Bankart repair in our department (50 arthroscopic suture anchor repairs, 5 arthroscopic screw fixations, and 26 open repairs). The 55 arthroscopic repairs comprised a consecutive cohort of patients treated by a single surgeon. Of the 50 patients in the suture anchor group, 45 (90%) were available for evaluation. At 82 months postoperatively, the mean Rowe score was 85.9 ± 20.5 points, the mean Western Ontario Shoulder Instability Index score was 89.4% ± 14.7%, and the mean Oxford Instability Score was 13.6 ± 5.4 points. Compared with the contralateral shoulder, all scores showed a significantly reduced outcome (P < .001, P < .001, and P < .001, respectively). A redislocation occurred in 3 patients (6.6%). Regarding satisfaction, 35 patients (78%) were very satisfied, 9 (20%) were satisfied, and 1 was partly satisfied. Overall, 95% of patients returned to any sporting activity after surgery. The number of sports disciplines (P < .001), duration (P 1⁄4 .005), level (P 1⁄4 .02), and risk category (P 1⁄4 .013) showed a significant reduction compared with the pretrauma condition. However, only 19% of patients reported that shoulder complaints were the reason for the redu

Published
2022

35. Histologische und molekularbiologische Analysen von mit MR-Kontrastmittel gelabelter mesenchymaler Stammzellen nach Implantation in osteochondrale Defekte im Kaninchenmodell

Author

Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Knolle, Percy A. (Prof. Dr.), Fleming, Megan Judith, Vogt, Stephan (Prof. Dr.), Vogt, Stephan (Prof. Dr.);Knolle, Percy A. (Prof. Dr.), and Fleming, Megan Judith

Abstract

In dieser Arbeit wurde der Einfluss einer Stammzellmarkierung mit dem MRT Eisenkontrastmittel Resovist® auf die Knorpeldefektregeneration untersucht sowie der histologische Nachweis der Markierung erbracht. In 36 Kaninchenfemora wurden allogene mesenchymale Stammzellen (MSC) in Fibrin–Zell–Implantaten in standardisierte Knorpeldefekte eingebracht. Die MSC waren mit eGFP für den histologischen Nachweis (n=36) sowie zur Hälfte mit Resovist® markiert (n=18). Die Auswertung erfolgte nach der 4., 12. und 24. Woche (je n=12). Der direkte Nachweis der Markierung gelang nach der 4. Woche uneingeschränkt, nach der 12. Woche nur in langsam heilenden Defekten. Weder makroskopisch noch mikroskopisch wurde gemäß ICRS Kriterien ein signifikanter Einfluss der Markierung nachgewiesen. Dies bestätigten auch die molekularbiologischen Analysen verschiedener Genexpressionsraten von typischen Differenzierungs , Dedifferenzierungs , sowie Entzündungsmarkern mittels PCR. Anhand dieser kumulativen Ergebnisse kann in Zukunft die Resovist® Markierung als in vivo Stammzellenmonitoring angewendet werden., Object of the study was to discern the influence of stem cell labelling with MRI-contrast media Resovist® on cartilage regeneration and histological label detection. Fibrin-Clots containing allogenic mesenchymal stem cells (MSC) were implanted in standardised cartilage defects in 36 rabbit femora. All MSC (n=36) were labelled lentivirally with eGFP for histological detection, in addition half of the MSC (n=18) were incubated with the MRI-contrast agent Resovist®. Ex vivo analysis was performed after 4, 12 and 24 weeks. After the 4th week histology showed both labels were detectable, after the 12th week detection was only possible in slowly healing defects. Microscopic and macroscopic evaluation of cartilage regeneration using ICRS criteria showed no significant differences through labelling. These findings were corroborated in molecular biological analysis of typical differentiation, dedifferentiation and inflammation markers using PCR. Henceforth, Resovist® can be used for in vivo stem cell monitoring.

Published
2022

37. Arthrosis

Author

Schuh, Reinhard, Hofstätter, Stefan, Trnka, Hans-Jörg, Schmitt, Andreas, Vogt, Stephan, Imhoff, Andreas B., editor, and Feucht, Matthias J., editor

Published
2015
Full Text
View/download PDF

41. Tendons

Author

Plath, Johannes E., Vogt, Stephan, Lenich, Andreas, Imhoff, Andreas B., editor, and Feucht, Matthias J., editor

Published
2015
Full Text
View/download PDF

43. PrOuD: Probabilistic Outlier Detection Solution for Time-Series Analysis of Real-World Photovoltaic Inverters.

Author

He, Yujiang, Huang, Zhixin, Vogt, Stephan, and Sick, Bernhard

Subjects
MONTE Carlo method, OUTLIER detection, TIME series analysis, TRUST, PROBLEM solving
Abstract

Anomaly detection methods applied to time series are mostly viewed as black boxes that solely provide a deterministic answer for the detected target. Without a convincing explanation, domain experts can hardly trust the detection results and must conduct further time-series diagnoses in real-world applications. To overcome this challenge, we mathematically analyzed the sources of anomalies and novelties in multivariate time series as well as their relationships from the perspective of Gaussian-distributed non-stationary noise. Furthermore, we proposed mathematical methods to generate artificial time series and synthetic anomalies, with the goal of solving the problem of it being difficult to train and evaluate models for real-world applications due to the lack of sufficient data. In addition, we designed Probabilistic Outlier Detection (PrOuD), which is a general solution to provide interpretable detection results to assist domain experts with time-series analysis. PrOuD can convert the predictive uncertainty of a time-series value from a trained model into the estimated uncertainty of the detected outlier through Monte Carlo estimation. The experimental results obtained on both artificial time series and real-world photovoltaic inverter data demonstrated that the proposed solution can detect emerging anomalies accurately and quickly. The implemented PrOuD demo case shows its potential to make the detection results of existing detection methods more convincing so that domain experts can more efficiently complete their tasks, such as time-series diagnosis and anomalous pattern clustering. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

45. Arthrose

Author

Schuh, Reinhard, Hofstätter, Stefan, Schmitt, Andreas, Vogt, Stephan, Trnka, Hans-Jörg, Imhoff, Andreas B., editor, and Feucht, Matthias J., editor

Published
2013
Full Text
View/download PDF

49. Sehnen

Author

Plath, Johannes, Lenich, Andreas, Vogt, Stephan, Imhoff, Andreas B., editor, and Feucht, Matthias J., editor

Published
2013
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results