HannIT AöR (Hannoversche Informationstechnologien – Anstalt öffentlichen Rechts) is the IT service provider for the Hannover region and about 30 other districts and municipalities in Lower Saxony. The security computer centre with distributed locations forms the system technical basis for the operation and further development of administrative procedures.
The Competence Centre Mobility and GIS is in charge of traffic information and planning systems, such as electronic timetable information or network availability maps. The services offered by hannIT AöR include

• the electronic timetable information in Lower Saxony and Bremen (EFA), among others for GVH (Großraum Verkehr Hannover) and Braunschweig ,
• the operation of the DELFI-Landesserver for Lower Saxony,
• the real-time connection of the public transport companies in the GVH
• multimodal routing (public transport, car, bicycle, pedestrian …)

In addition, hannIT as a service provider is significantly involved in projects such as

• Connect GmbH (data integration for Lower Saxony and Bremen),
• DELFI e.V. (Germany-wide timetable information),
• EU-Spirit (Association of European directory assistance systems)

In the D4UM project hannIT AöR 

• ensures that the platform (hardware and software) is available
• coordinates the provision of SW solutions,
• collaborates on data protection and security rules
• creates the basis for the integration of IV real-time data and historical charts into the EFA timetable information (Prototype)

• ATLAS: Adaptive Attribute-Aware Post-Hoc Alignment of Knowledge Graph Embeddings. Ekinci Birol, Duygu; Kouagou, N’Dah Jean; Sherif, Mohamed Ahmed; Ngonga Ngomo, Axel-Cyrille (2026).
• RANGEFC: Interval-Aware Temporal Fact Checking for Knowledge Graph. Qamar, Abdullah; Qudus, Umair; Röder, Michael; Ngomo, Axel-Cyrille Ngonga U. Sattler, A. Hogan, M. Keet, V. Presutti, J. P. A. Almeida, H. Takeda, P. Monnin, G. Pirrò, C. d’Amato (eds.) (2026).
• Civil Engineering and Urban Planning: An International Journal (CiVEJ). Domingos, Pedro; Pazzani, Michael J. (C. Engineering; U. P. A. I. J. (CiVEJ), eds.) (2026). 29(2-3) 103–130.

  . The simple Bayesian classifier is known to be optimal when attributes are independent given the class, but the question of whether other sufficient conditions for its optimality exist has so far not been explored. Empirical results showing that it performs surprisingly well in many domains containing clear attribute dependences suggest that the answer to this question may be positive. This article shows that, although the Bayesian classifier's probability estimates are only optimal under...
• OPTIMIZING DATA INTEROPERABILITY IN AGILE ORGANIZATIONS:INTEGRATING NONAKA’S BA AND HABERMAS’S DELIBERATION FOR TRUST-BASED KNOWLEDGE MANAGEMENT. 1, Marina Özdemir; and Cihangir Deniz Özdemir 2 (M. Özdemir 1; and Cihangir Deniz Özdemir 2, eds.) (2026). 6(4) 776–785.

  Agile methodologies have transformed organizational management by prioritizing team autonomy and iterative learning cycles. However, these approaches often lack structured mechanisms for knowledge retention and interoperability, leading to fragmented decision-making, information silos, and strategic misalignment. This study proposes an alternative approach to knowledge management in Agile environments by integrating Ikujiro Nonaka and Hirotaka Takeuchi’s theory of knowledge creation— specifically the concept of Ba, a shared space where knowledge is created and validated—with Jürgen Habermas’s Theory of Communicative Action, which emphasizes deliberation as the foundation for trust and legitimacy in organizational decision-making. To operationalize this integration, we propose the Deliberative Permeability Metric (DPM), a diagnostic tool that evaluates knowledge flow and the deliberative foundation of organizational decisions, and the Communicative Rationality Cycle (CRC), a structured feedback model that extends the DPM, ensuring long-term adaptability and data governance. This model was applied at Livelo, a Brazilian loyalty program company, demonstrating that structured deliberation improves operational efficiency and reduces knowledge fragmentation. The findings indicate that institutionalizing deliberative processes strengthens knowledge interoperability, fostering a more resilient and adaptive approach to data governance in complex organizations.
• The International Journal of Database Management Systems (IJDMS). Journal, Ijdms (I. Journal, ed.) (2026).
• AdaCLIP: Adapting CLIP with Hybrid Learnable Prompts for Zero-Shot Anomaly Detection. Cao, Yunkang; Zhang, Jiangning; Frittoli, Luca; Cheng, Yuqi; Shen, Weiming; Boracchi, Giacomo A. Leonardis, E. Ricci, S. Roth, O. Russakovsky, T. Sattler, G. Varol (eds.) (2025). 55–72.

  Zero-shot anomaly detection (ZSAD) targets the identification of anomalies within images from arbitrary novel categories. This study introduces AdaCLIP for the ZSAD task, leveraging a pre-trained vision-language model (VLM), CLIP. AdaCLIP incorporates learnable prompts into CLIP and optimizes them through training on auxiliary annotated anomaly detection data. Two types of learnable prompts are proposed: static and dynamic. Static prompts are shared across all images, serving to preliminarily adapt CLIP for ZSAD. In contrast, dynamic prompts are generated for each test image, providing CLIP with dynamic adaptation capabilities. The combination of static and dynamic prompts is referred to as hybrid prompts, and yields enhanced ZSAD performance. Extensive experiments conducted across 14 real-world anomaly detection datasets from industrial and medical domains indicate that AdaCLIP outperforms other ZSAD methods and can generalize better to different categories and even domains. Finally, our analysis highlights the importance of diverse auxiliary data and optimized prompts for enhanced generalization capacity. Code is available at https://github.com/caoyunkang/AdaCLIP.
• Beyond the usual: review of rare cases of primary breast mesenchymal sarcoma from a tertiary institution. Chowdhury, Zachariah; Makker, Isha; Hayat, Sadaf; Paul, Paramita; Patne, Shashikant; Dhal, Ipista (2025). 13(6) 7–14.

  Introduction: Primary breast sarcomas (PBSs) account for <1% of all breast cancer cases, typically affecting patients aged 55–59 years. They are difficult to differentiate from other types of breast cancer based on the clinical and imaging results and are associated with a poor prognosis. It is important to recognize them owing to differences in treatment options. We here present four cases of rare PBSs diagnosed at our center. Aims and Objectives: The purpose of presenting this case series is to familiarize the pathologist with these rare entities as they clinically mimic epithelial malignancy of the breast but are heterogeneous and totally follow a different clinical course and treatment protocol. Materials and Methods: A collection of four rare breast mesenchymal tumors is presented. Clinical details of patients, including follow–up, were obtained from the electronic medical record of hospital. Results/Cases: A total of four cases were identified. All were female with age group of presentation being 14–71 years with mean age being 44 years. The tumor size ranged from 1.5 cm to 20 cm. Microscopically, a diagnosis of Rhabdomyosarcoma was rendered in one case, one was diagnosed as malignant peripheral nerve sheath tumors, one as osteosarcoma, and the last as atypical cartilaginous tumor. A varied panel of immunohistochemistry (IHC) markers was used to reach the final diagnosis, including SMA, Desmin, CD34, S-100, H3K27Me3, Myogenin, MyoD1, and SATB2 among others. Most common differential in all these cases was a malignant phyllodes tumor. Conclusion: This case series describes unique cases of PBSs. Histopathological analysis has great importance in the diagnosis of these tumors, along with an IHC panel that is valuable in excluding phyllodes tumors and differentiating among the various sarcomas. In our setting, IHC was a major input in diagnosis. This study supports and builds on global data on rare breast neoplasms, while offering valuable regional insights.
• Entanglement as a More Fundamental Notion of Information than Shannon and von Neumann Entropy. Barry, Michael W. (2025).
• Automating IMRT planning for cervical cancer using dimension-scaled prior-based vanilla bayesian optimization. Saini, Vinay; Barman, Sanjay; Pradhan, Satyajit; Shinghal, Abhishek; Kapoor, Ankita; Sharma, Neeraj (2025). 70(23) 235034-.

  Objective. Manual inverse planning for radiation therapy is labor-intensive and often prone to inconsistent plan quality due to multiple adjustable planning hyperparameters, varying planner experience, and differing time constraints. Automated treatment planning offers a solution to these challenges. The present study investigates the effectiveness of dimension-scaled prior (DSP) Vanilla Bayesian optimization (BO) with a log-expected improvement (logEI) acquisition function in automating intensity-modulated radiation therapy (IMRT) planning for cancer cervix (CaCx) in high-dimensional settings. Approach. A Python-based auto-optimization script utilizing DSP Vanilla BO with logEI was employed to iteratively optimize the planning hyperparameters, including dose objectives and their corresponding weights for CaCx case IMRT plans on the Varian Eclipse treatment planning system (TPS) v18.0. This approach was assessed in 30 retrospectively selected pelvic node-positive CaCx cases, and the dosimetric parameters, based on EMBRACE-II protocol, were compared with plans generated from manual, Sparse Axis-Aligned Subspace BO (SAASBO), and stopping criterion-based DSP Vanilla BO. Main results. DSP Vanilla BO plans demonstrated superior dose conformity ( & ) and organs at risk (OAR) sparing (, , , , , , and ) compared to manual planning with significant improvements (p < 0.05), while maintaining adequate clinical target coverage for CTVN, PTV55, ITV45, and PTV45. Compared to SAASBO, DSP Vanilla BO achieved comparable dosimetric quality but with less computation time (∼94 vs 360 min). The addition of stopping criteria further reduced the optimization time to ∼44 min while maintaining a plan quality comparable to manual planning. Significance. The study demonstrated that DSP Vanilla BO automated plans achieved comparable target coverage, with an improvement in OAR sparing, compared to manual plans. This highlights its effectiveness as an efficient, data-independent method for automating IMRT planning, which can be easily integrated into a TPS and benefit clinics with limited resources.
• Association of chewing tobacco and the risk of breast cancer in Indian women: a multicentre case-control study. Moirangthem, Romi; Manjrekar, Ankita; Pullat, Gayathri B.; Golapkar, Shruti Vishwas; Margale, Ruchita Sahdev; Sagare, Sakshi; Pathak, Ruchi; Pradhan, Satyajit; Krishnatreya, Manigreeva; Panse, Nandkumar; Dixit, Rajesh; Soerjomataram, Isabelle; Gupta, Sudeep; Mhatre, Sharayu (2025).

  Background: Even though breast cancer (BC) is the most common female cancer worldwide, the role of tobacco, specifically chewing tobacco in development of BC has not been widely studied. This study aimed to assess this association. Methods: A multicentre hospital-based case-control study was utilised. 2559 histopathologically confirmed BC cases, and 2239 visitor controls were included. Self-reported information was collected regarding tobacco consumption and other potential confounders. Logistic regression model was used to calculate odds ratio (OR) and its 95% confidence interval (CI), after adjusting for age, current residential region, education, various reproductive factors, BMI and history of benign breast lump. Attributable fraction (AF) and population attributable fraction (PAF) of BC due to chewing tobacco was also calculated. Findings: An increased risk of BC was observed in women who ever used chewing tobacco (OR:1.20, 95% CI:1.01-1.42) as compared to those who never consumed tobacco (smoking and chewing), consistent across all subtypes of BC. Dose-response relation was observed for duration of tobacco chewing (OR>25 years: 1.36, 95% CI: 1.03-1.80). Women who initiated chewing tobacco at <20 years, before their first full-term pregnancy (FFTP) had more risk. Observed association was consistent even after stratification on menopausal status. The AF of BC due to tobacco chewing in our study was calculated to be approximately 3%, while PAF for India was about 2%. Interpretation: Our study suggests that chewing tobacco is associated with an increased risk of BC, for all subtypes. This is particularly true when the duration of exposure is higher and, exposure begins before FFTP. This highlights the need to target tobacco control policies to smokeless tobacco along with smoking, thus reducing the burden of BC to some extent.FundingThis work was supported by the Tata Memorial Centre, Mumbai (Intramural fund).
• A Multi-Head Model for Continual Learning via Out-of-Distribution Replay. Kim, Gyuhak; Ke, Zixuan; Liu, Bing (2022).
• 線形モデルを用いた特徴選択と非線形モデル解釈による特徴の重要度の比較と考察. 武子大輔 (2022).
• Switch transformers: scaling to trillion parameter models with simple and efficient sparsity. Fedus, William; Zoph, Barret; Shazeer, Noam (2022). 23(1)

  In deep learning, models typically reuse the same parameters for all inputs. Mixture of Experts (MoE) models defy this and instead select different parameters for each incoming example. The result is a sparsely-activated model--with an outrageous number of parameters--but a constant computational cost. However, despite several notable successes of MoE, widespread adoption has been hindered by complexity, communication costs, and training instability. We address these with the introduction of the Switch Transformer. We simplify the MoE routing algorithm and design intuitive improved models with reduced communication and computational costs. Our proposed training techniques mitigate the instabilities, and we show large sparse models may be trained, for the first time, with lower precision (bfloat16) formats. We design models based off T5-Base and T5-Large (Raffel et al., 2019) to obtain up to 7x increases in pre-training speed with the same computational resources. These improvements extend into multilingual settings where we measure gains over the mT5-Base version across all 101 languages. Finally, we advance the current scale of language models by pre-training up to trillion parameter models on the "Colossal Clean Crawled Corpus", and achieve a 4x speedup over the T5-XXL model.
• Fragebogen zur Erfassung der autonomen und kontrollierten Motivation zur (Nicht-)Beschäftigung mit der beruflichen Zukunft von Jugendlichen mit und ohne sonderpädagogische Förderbedarfe. Kashikar, L. & Grosche (2022).
• Fourier Analysis Körner, T. W. in Cambridge Mathematical Library (2022). Cambridge University Press.
• A patient-centric dataset of images and metadata for identifying melanomas using clinical context. Rotemberg, Veronica; Kurtansky, Nicholas; Betz-Stablein, Brigid; Caffery, Liam; Chousakos, Emmanouil; Codella, Noel; Combalia, Marc; Dusza, Stephen; Guitera, Pascale; Gutman, David; Halpern, Allan; Helba, Brian; Kittler, Harald; Kose, Kivanc; Langer, Steve; Lioprys, Konstantinos; Malvehy, Josep; Musthaq, Shenara; Nanda, Jabpani; Reiter, Ofer; Shih, George; Stratigos, Alexander; Tschandl, Philipp; Weber, Jochen; Soyer, H. Peter (2021). 8(1)
• A Unifying Review of Deep and Shallow Anomaly Detection. Ruff, Lukas; Kauffmann, Jacob R.; Vandermeulen, Robert A.; Montavon, Grégoire; Samek, Wojciech; Kloft, Marius; Dietterich, Thomas G.; Müller, Klaus-Robert (2021). 109(5) 756–795.

  Deep learning approaches to anomaly detection (AD) have recently improved the state of the art in detection performance on complex data sets, such as large collections of images or text. These results have sparked a renewed interest in the AD problem and led to the introduction of a great variety of new methods. With the emergence of numerous such methods, including approaches based on generative models, one-class classification, and reconstruction, there is a growing need to bring methods of this field into a systematic and unified perspective. In this review, we aim to identify the common underlying principles and the assumptions that are often made implicitly by various methods. In particular, we draw connections between classic “shallow” and novel deep approaches and show how this relation might cross-fertilize or extend both directions. We further provide an empirical assessment of major existing methods that are enriched by the use of recent explainability techniques and present specific worked-through examples together with practical advice. Finally, we outline critical open challenges and identify specific paths for future research in AD.
• Pneumonia detection in chest X-ray images using an ensemble of deep learning models. Kundu, Rohit; Das, Ritacheta; Geem, Zong Woo; Han, Gi-Tae; Sarkar, Ram (2021). 16(9) 1–29.

  Pneumonia is a respiratory infection caused by bacteria or viruses; it affects many individuals, especially in developing and underdeveloped nations, where high levels of pollution, unhygienic living conditions, and overcrowding are relatively common, together with inadequate medical infrastructure. Pneumonia causes pleural effusion, a condition in which fluids fill the lung, causing respiratory difficulty. Early diagnosis of pneumonia is crucial to ensure curative treatment and increase survival rates. Chest X-ray imaging is the most frequently used method for diagnosing pneumonia. However, the examination of chest X-rays is a challenging task and is prone to subjective variability. In this study, we developed a computer-aided diagnosis system for automatic pneumonia detection using chest X-ray images. We employed deep transfer learning to handle the scarcity of available data and designed an ensemble of three convolutional neural network models: GoogLeNet, ResNet-18, and DenseNet-121. A weighted average ensemble technique was adopted, wherein the weights assigned to the base learners were determined using a novel approach. The scores of four standard evaluation metrics, precision, recall, f1-score, and the area under the curve, are fused to form the weight vector, which in studies in the literature was frequently set experimentally, a method that is prone to error. The proposed approach was evaluated on two publicly available pneumonia X-ray datasets, provided by Kermany et al. and the Radiological Society of North America (RSNA), respectively, using a five-fold cross-validation scheme. The proposed method achieved accuracy rates of 98.81% and 86.85% and sensitivity rates of 98.80% and 87.02% on the Kermany and RSNA datasets, respectively. The results were superior to those of state-of-the-art methods and our method performed better than the widely used ensemble techniques. Statistical analyses on the datasets using McNemar’s and ANOVA tests showed the robustness of the approach. The codes for the proposed work are available at https://github.com/Rohit-Kundu/Ensemble-Pneumonia-Detection.
• 機械学習を解釈する技術: 予測力と説明力を両立する実践テクニック 森下光之助 (2021). 技術評論社.
• Learning Placeholders for Open-Set Recognition. Zhou, Da-Wei; Ye, Han-Jia; Zhan, De-Chuan (2021).
• Deep Learning for Anomaly Detection: A Review. Pang, Guansong; Shen, Chunhua; Cao, Longbing; Hengel, Anton Van Den (2021). 54(2)

  Anomaly detection, a.k.a. outlier detection or novelty detection, has been a lasting yet active research area in various research communities for several decades. There are still some unique problem complexities and challenges that require advanced approaches. In recent years, deep learning enabled anomaly detection, i.e., deep anomaly detection, has emerged as a critical direction. This article surveys the research of deep anomaly detection with a comprehensive taxonomy, covering advancements in 3 high-level categories and 11 fine-grained categories of the methods. We review their key intuitions, objective functions, underlying assumptions, advantages, and disadvantages and discuss how they address the aforementioned challenges. We further discuss a set of possible future opportunities and new perspectives on addressing the challenges.
• 勾配ブースティングを用いたCOVID-19感染症の重症度判別と要因の分析. 長田卓馬 (2021).
• Dynamic Spatial-Temporal Representation Learning for Traffic Flow Prediction. Liu, Lingbo; Zhen, Jiajie; Li, Guanbin; Zhan, Geng; He, Zhaocheng; Du, Bowen; Lin, Liang (2021). 22(11) 7169–7183.

  As a crucial component in intelligent transportation systems, traffic flow prediction has recently attracted widespread research interest in the field of artificial intelligence (AI) with the increasing availability of massive traffic mobility data. Its key challenge lies in how to integrate diverse factors (such as temporal rules and spatial dependencies) to infer the evolution trend of traffic flow. To address this problem, we propose a unified neural network called Attentive Traffic Flow Machine (ATFM), which can effectively learn the spatial-temporal feature representations of traffic flow with an attention mechanism. In particular, our ATFM is composed of two progressive Convolutional Long Short-Term Memory (ConvLSTM [1]) units connected with a convolutional layer. Specifically, the first ConvLSTM unit takes normal traffic flow features as input and generates a hidden state at each time-step, which is further fed into the connected convolutional layer for spatial attention map inference. The second ConvLSTM unit aims at learning the dynamic spatial-temporal representations from the attentionally weighted traffic flow features. Further, we develop two deep learning frameworks based on ATFM to predict citywide short-term/long-term traffic flow by adaptively incorporating the sequential and periodic data as well as other external influences. Extensive experiments on two standard benchmarks well demonstrate the superiority of the proposed method for traffic flow prediction. Moreover, to verify the generalization of our method, we also apply the customized framework to forecast the passenger pickup/dropoff demands in traffic prediction and show its superior performance. Our code and data are available at https://github.com/liulingbo918/ATFM.
• AI/深層学習を損傷・異常検知に使用するには　～方法論の整理と適用事例. 野村泰稔 (2020).
• Out-of-distribution detection for high dimensional data using low dimensional representation with deep generative models. OKAMOTO, Hirono; SUZUKI, Masahiro; MATSUO, Yutaka (2020). JSAI2020 1J5GS203–1J5GS203.
• Generalizing from a Few Examples: A Survey on Few-shot Learning. Wang, Yaqing; Yao, Quanming; Kwok, James T.; Ni, Lionel M. (2020). 53(3) 63:1–63:34.
• Generalizing from a Few Examples. Wang, Yaqing; Yao, Quanming; Kwok, James T.; Ni, Lionel M. (2020). 53(3) 1–34.
• Unsupervised anomaly detection using Exact Rate Distorion Auto Encoder. NAKANISHI, Hitoshi; SUZUKI, Masahiro; MATSUO, Yutaka (2020). JSAI2020 4Rin186–4Rin186.
• Predicting spikes in cab ride time series using a deep neural network. YAMAMOTO, Ayaka; YANG, Bojian; ANDO, Shin (2020). JSAI2020 4Rin188–4Rin188.
• オートエンコーダとは？事前学習の仕組み・現在の活用方法を解説！！. ITトレンド (2020).
• 決定木モデルの解釈におけるSHAP値の有用性の検証. 秀穂吉田; 優樹田嶋; 優作今井 (2020). JSAI2020 3E5GS204–3E5GS204.
• Research Trends in XAI(eXplainable AI) Technologies. Egi, Masashi (2020). 34(1) 20–27.
• Verification of Usefulness of SHAP values in Interpretation of Decision Tree Models. YOSHIDA, Shuho; TAJIMA, Yuki; IMAI, Yusaku (2020). JSAI2020 3E5GS204–3E5GS204.
• AI・機械学習のデータセット辞典：MNIST：手書き数字の画像データセット. 一色政彦 (2020).
• Multi-Class Data Description for Out-of-distribution Detection. Lee, Dongha; Yu, Sehun; Yu, Hwanjo (2020).
• MathWorkstextregistered　「異常検知」. Inc., The MathWorkstextregistered (2020).
• Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization. Selvaraju, Ramprasaath R.; Cogswell, Michael; Das, Abhishek; Vedantam, Ramakrishna; Parikh, Devi; Batra, Dhruv (2020). 128(2) 336–359.
• On Out-of-Distribution Detection Algorithms with Deep Neural Skin Cancer Classifiers. Pacheco, André G. C.; Sastry, Chandramouli Shama; Trappenberg, Thomas; Oore, Sageev; Krohling, Renato A. (2020). 3152–3161.
• 多変量時系列を入力とする金価格時系列予測. 石森有紗; 坂本茉菜 (2020).
• Analyzing and Improving the Image Quality of StyleGAN. Karras, Tero; Laine, Samuli; Aittala, Miika; Hellsten, Janne; Lehtinen, Jaakko; Aila, Timo (2020). 8107–8116.
• Deep Semi-Supervised Anomaly Detection. Ruff, Lukas; Vandermeulen, Robert A.; Görnitz, Nico; Binder, Alexander; Müller, Emmanuel; Müller, Klaus-Robert; Kloft, Marius (2020).
• Deep Semi-Supervised Anomaly Detection. Ruff, Lukas; Vandermeulen, Robert A.; Görnitz, Nico; Binder, Alexander; Müller, Emmanuel; Müller, Klaus-Robert; Kloft, Marius (2020).
• Deep Variational Semi-Supervised Novelty Detection. Daniel, Tal; Kurutach, Thanard; Tamar, Aviv (2019). abs/1911.04971
• Deep Learning for Anomaly Detection: A Survey. Chalapathy, Raghavendra; Chawla, Sanjay (2019).
• 東京大学のデータサイエンティスト育成講座 塚本邦尊; 山田典一; 大澤文孝; 中山浩太郎; 松尾豊 (2019). マイナビ出版.
• Real-time Convolutional Neural Networks for emotion and gender classification. Valdenegro-Toro, Matias; Arriaga, Octavio; Plöger, Paul (2019).
• Deep Semi-Supervised Anomaly Detection. Ruff, Lukas; Vandermeulen, Robert A.; Görnitz, Nico; Binder, Alexander; Müller, Emmanuel; Müller, Klaus-Robert; Kloft, Marius (2019). abs/1906.02694
• 第６回不均衡データ学習(Learning from Imbalanced Data)を学ぶ(3). 吉村皐亮 (2019).
• Image Anomaly Detection with Generative Adversarial Networks. Deecke, Lucas; Vandermeulen, Robert; Ruff, Lukas; Mandt, Stephan; Kloft, Marius M. Berlingerio, F. Bonchi, T. Gärtner, N. Hurley, G. Ifrim (eds.) (2019). 3–17.

  Many anomaly detection methods exist that perform well on low-dimensional problems however there is a notable lack of effective methods for high-dimensional spaces, such as images. Inspired by recent successes in deep learning we propose a novel approach to anomaly detection using generative adversarial networks. Given a sample under consideration, our method is based on searching for a good representation of that sample in the latent space of the generator; if such a representation is not found, the sample is deemed anomalous. We achieve state-of-the-art performance on standard image benchmark datasets and visual inspection of the most anomalous samples reveals that our method does indeed return anomalies.
• Real-time Convolutional Neural Networks for emotion and gender classification. Valdenegro-Toro, Matias; Arriaga, Octavio; Plöger, Paul (2019).
• Thoth children 「ROC曲線」. children, Thoth (2019).
• Labeled Optical Coherence Tomography (OCT) and Chest X-Ray Images for Classification. Kermany, Daniel (2018).
• Adversarially Learned Anomaly Detection. Zenati, Houssam; Romain, Manon; Foo, Chuan-Sheng; Lecouat, Bruno; Chandrasekhar, Vijay (2018). 727–736.
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. Devlin, Jacob; Chang, Ming-Wei; Lee, Kenton; Toutanova, Kristina (2018). abs/1810.04805
• Efficient GAN-Based Anomaly Detection. Zenati, Houssam; Foo, Chuan Sheng; Lecouat, Bruno; Manek, Gaurav; Chandrasekhar, Vijay Ramaseshan (2018). abs/1802.06222
• Kitsune: An Ensemble of Autoencoders for Online Network Intrusion Detection. Mirsky, Yisroel; Doitshman, Tomer; Elovici, Yuval; Shabtai, Asaf (2018).

  Neural networks have become an increasingly popular solution for network intrusion detection systems (NIDS). Their capability of learning complex patterns and behaviors make them a suitable solution for differentiating between normal traffic and network attacks. However, a drawback of neural networks is the amount of resources needed to train them. Many network gateways and routers devices, which could potentially host an NIDS, simply do not have the memory or processing power to train and sometimes even execute such models. More importantly, the existing neural network solutions are trained in a supervised manner. Meaning that an expert must label the network traffic and update the model manually from time to time. In this paper, we present Kitsune: a plug and play NIDS which can learn to detect attacks on the local network, without supervision, and in an efficient online manner. Kitsune's core algorithm (KitNET) uses an ensemble of neural networks called autoencoders to collectively differentiate between normal and abnormal traffic patterns. KitNET is supported by a feature extraction framework which efficiently tracks the patterns of every network channel. Our evaluations show that Kitsune can detect various attacks with a performance comparable to offline anomaly detectors, even on a Raspberry PI. This demonstrates that Kitsune can be a practical and economic NIDS.
• Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning. Kermany, Daniel S.; Goldbaum, Michael; Cai, Wenjia; Valentim, Carolina C.S.; Liang, Huiying; Baxter, Sally L.; McKeown, Alex; Yang, Ge; Wu, Xiaokang; Yan, Fangbing; Dong, Justin; Prasadha, Made K.; Pei, Jacqueline; Ting, Magdalene Y.L.; Zhu, Jie; Li, Christina; Hewett, Sierra; Dong, Jason; Ziyar, Ian; Shi, Alexander; Zhang, Runze; Zheng, Lianghong; Hou, Rui; Shi, William; Fu, Xin; Duan, Yaou; Huu, Viet A.N.; Wen, Cindy; Zhang, Edward D.; Zhang, Charlotte L.; Li, Oulan; Wang, Xiaobo; Singer, Michael A.; Sun, Xiaodong; Xu, Jie; Tafreshi, Ali; Lewis, M. Anthony; Xia, Huimin; Zhang, Kang (2018). 172(5) 1122–1131.e9.
• GANomaly: Semi-Supervised Anomaly Detection via Adversarial Training. Akcay, Samet; Atapour-Abarghouei, Amir; Breckon, Toby P. (2018).
• Deep One-Class Classification. Ruff, Lukas; Görnitz, Nico; Deecke, Lucas; Siddiqui, Shoaib Ahmed; Vandermeulen, Robert A.; Binder, Alexander; Müller, Emmanuel; Kloft, Marius in Proceedings of Machine Learning Research, J. G. Dy, A. Krause (eds.) (2018). (Vol. 80) 4390–4399.
• Deep One-Class Classification. Ruff, Lukas; Vandermeulen, Robert; Goernitz, Nico; Deecke, Lucas; Siddiqui, Shoaib Ahmed; Binder, Alexander; Müller, Emmanuel; Kloft, Marius in Proceedings of Machine Learning Research, J. Dy, A. Krause (eds.) (2018). (Vol. 80) 4393–4402.

  Despite the great advances made by deep learning in many machine learning problems, there is a relative dearth of deep learning approaches for anomaly detection. Those approaches which do exist involve networks trained to perform a task other than anomaly detection, namely generative models or compression, which are in turn adapted for use in anomaly detection; they are not trained on an anomaly detection based objective. In this paper we introduce a new anomaly detection method—Deep Support Vector Data Description—, which is trained on an anomaly detection based objective. The adaptation to the deep regime necessitates that our neural network and training procedure satisfy certain properties, which we demonstrate theoretically. We show the effectiveness of our method on MNIST and CIFAR-10 image benchmark datasets as well as on the detection of adversarial examples of GTSRB stop signs.
• Music Transformer. Huang, Cheng-Zhi Anna; Vaswani, Ashish; Uszkoreit, Jakob; Shazeer, Noam; Simon, Ian; Hawthorne, Curtis; Dai, Andrew M.; Hoffman, Matthew D.; Dinculescu, Monica; Eck, Douglas (2018).

  Music relies heavily on repetition to build structure and meaning. Self-reference occurs on multiple timescales, from motifs to phrases to reusing of entire sections of music, such as in pieces with ABA structure. The Transformer (Vaswani et al., 2017), a sequence model based on self-attention, has achieved compelling results in many generation tasks that require maintaining long-range coherence. This suggests that self-attention might also be well-suited to modeling music. In musical composition and performance, however, relative timing is critically important. Existing approaches for representing relative positional information in the Transformer modulate attention based on pairwise distance (Shaw et al., 2018). This is impractical for long sequences such as musical compositions since their memory complexity for intermediate relative information is quadratic in the sequence length. We propose an algorithm that reduces their intermediate memory requirement to linear in the sequence length. This enables us to demonstrate that a Transformer with our modified relative attention mechanism can generate minute-long compositions (thousands of steps, four times the length modeled in Oore et al., 2018) with compelling structure, generate continuations that coherently elaborate on a given motif, and in a seq2seq setup generate accompaniments conditioned on melodies. We evaluate the Transformer with our relative attention mechanism on two datasets, JSB Chorales and Piano-e-Competition, and obtain state-of-the-art results on the latter.
• LSTM. 正孝徳丸 (2018). 30巻4号 223.
• Deep Anomaly Detection Using Geometric Transformations. Golan, Izhak; El-Yaniv, Ran in NIPS’18 (2018). 9781–9791.

  We consider the problem of anomaly detection in images, and present a new detection technique. Given a sample of images, all known to belong to a "normal" class (e.g., dogs), we show how to train a deep neural model that can detect out-of-distribution images (i.e., non-dog objects). The main idea behind our scheme is to train a multi-class model to discriminate between dozens of geometric transformations applied on all the given images. The auxiliary expertise learned by the model generates feature detectors that effectively identify, at test time, anomalous images based on the softmax activation statistics of the model when applied on transformed images. We present extensive experiments using the proposed detector, which indicate that our technique consistently improves all known algorithms by a wide margin.
• オートエンコーダによる低次元化と可視化. 範泰尾亦 (2018). Vol.38 No.151 9–13.
• Deep Over-sampling Framework for Classifying Imbalanced Data. Ando, Shin; Huang, Chun Yuan M. Ceci, J. Hollmén, L. Todorovski, C. Vens, S. Dvzeroski (eds.) (2017). 770–785.

  Class imbalance is a challenging issue in practical classification problems for deep learning models as well as traditional models. Traditionally successful countermeasures such as synthetic over-sampling have had limited success with complex, structured data handled by deep learning models. In this paper, we propose Deep Over-sampling (DOS), a framework for extending the synthetic over-sampling method to the deep feature space acquired by a convolutional neural network (CNN). Its key feature is an explicit, supervised representation learning, for which the training data presents each raw input sample with a synthetic embedding target in the deep feature space, which is sampled from the linear subspace of in-class neighbors. We implement an iterative process of training the CNN and updating the targets, which induces smaller in-class variance among the embeddings, to increase the discriminative power of the deep representation. We present an empirical study using public benchmarks, which shows that the DOS framework not only counteracts class imbalance better than the existing method, but also improves the performance of the CNN in the standard, balanced settings.
• LSTMを用いた株価変動予測. 藤五郎松井; 智哉汐月 (2017). JSAI2017 2D3OS19a2–2D3OS19a2.
• Xception: Deep learning with depthwise separable convolutions. Chollet, François (2017). 1251–1258.
• Unsupervised Anomaly Detection with Generative Adversarial Networks to Guide Marker Discovery. Schlegl, Thomas; Seeböck, Philipp; Waldstein, Sebastian M.; Schmidt-Erfurth, Ursula; Langs, Georg (2017). abs/1703.05921
• Proximal Policy Optimization Algorithms. Schulman, John; Wolski, Filip; Dhariwal, Prafulla; Radford, Alec; Klimov, Oleg (2017). abs/1707.06347

  We propose a new family of policy gradient methods for reinforcement learning, which alternate between sampling data through interaction with the environment, and optimizing a "surrogate" objective function using stochastic gradient ascent. Whereas standard policy gradient methods perform one gradient update per data sample, we propose a novel objective function that enables multiple epochs of minibatch updates. The new methods, which we call proximal policy optimization (PPO), have some of the benefits of trust region policy optimization (TRPO), but they are much simpler to implement, more general, and have better sample complexity (empirically). Our experiments test PPO on a collection of benchmark tasks, including simulated robotic locomotion and Atari game playing, and we show that PPO outperforms other online policy gradient methods, and overall strikes a favorable balance between sample complexity, simplicity, and wall-time.
• LSTMを用いた株価変動予測. 藤五郎松井; 智也汐月 (2017). pp.2D3OS19a2, 2017
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms. Xiao, Han; Rasul, Kashif; Vollgraf, Roland (2017).

  We present Fashion-MNIST, a new dataset comprising of 28x28 grayscale images of 70,000 fashion products from 10 categories, with 7,000 images per category. The training set has 60,000 images and the test set has 10,000 images. Fashion-MNIST is intended to serve as a direct drop-in replacement for the original MNIST dataset for benchmarking machine learning algorithms, as it shares the same image size, data format and the structure of training and testing splits. The dataset is freely available at https://github.com/zalandoresearch/fashion-mnist
• LightGBM: A Highly Efficient Gradient Boosting Decision Tree. Ke, Guolin; Meng, Qi; Finley, Thomas; Wang, Taifeng; Chen, Wei; Ma, Weidong; Ye, Qiwei; Liu, Tie-Yan I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, R. Garnett (eds.) (2017). 3146–3154.
• Unsupervised Anomaly Detection with Generative Adversarial Networks to Guide Marker Discovery. Schlegl, Thomas; Seeböck, Philipp; Waldstein, Sebastian M.; Schmidt-Erfurth, Ursula; Langs, Georg in Lecture Notes in Computer Science, M. Niethammer, M. Styner, S. R. Aylward, H. Zhu, I. Oguz, P.-T. Yap, D. Shen (eds.) (2017). (Vol. 10265) 146–157.
• A Unified Approach to Interpreting Model Predictions. Lundberg, Scott M; Lee, Su-In I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, R. Garnett (eds.) (2017). 4765–4774.
• LSTMを用いた株価変動予測. 藤五郎松井; 智哉汐月 (2017). (Vol. JSAI2017) 2D3OS19a2–2D3OS19a2.
• Random Forests: Fundamentals and Recent Trends. Habe, Hitoshi (2016). 70(9) 788–791.
• XGBoost. Chen, Tianqi; Guestrin, Carlos (2016).
• Frontiers in Distributed Representations for Natural Language Processing(<Special Issue>Research Frontiers in Neural Network). OKAZAKI, Naoaki (2016). 31(2) 189–201.
• Deep Learning Goodfellow, Ian; Bengio, Yoshua; Courville, Aaron (2016).
• Striving for Simplicity: The All Convolutional Net. Springenberg, Jost Tobias; Dosovitskiy, Alexey; Brox, Thomas; Riedmiller, Martin A. Y. Bengio, Y. LeCun (eds.) (2015).
• Deep learning and the information bottleneck principle. Tishby, Naftali; Zaslavsky, Noga (2015). 1–5.

  Deep Neural Networks (DNNs) are analyzed via the theoretical framework of the information bottleneck (IB) principle. We first show that any DNN can be quantified by the mutual information between the layers and the input and output variables. Using this representation we can calculate the optimal information theoretic limits of the DNN and obtain finite sample generalization bounds. The advantage of getting closer to the theoretical limit is quantifiable both by the generalization bound and by the network's simplicity. We argue that both the optimal architecture, number of layers and features/connections at each layer, are related to the bifurcation points of the information bottleneck tradeoff, namely, relevant compression of the input layer with respect to the output layer. The hierarchical representations at the layered network naturally correspond to the structural phase transitions along the information curve. We believe that this new insight can lead to new optimality bounds and deep learning algorithms.
• Example-Dependent Cost-Sensitive Classification with Applications in Financial Risk Modeling and Marketing Analytics. Technical Report (PhD dissertation), Bahnsen, Alejandro Correa (2015).
• Visualizing and Understanding Convolutional Networks. Zeiler, Matthew D.; Fergus, Rob in Lecture Notes in Computer Science, D. J. Fleet, T. Pajdla, B. Schiele, T. Tuytelaars (eds.) (2014). (Vol. 8689) 818–833.
• Distributed Representations of Sentences and Documents. Le, Quoc; Mikolov, Tomas in Proceedings of Machine Learning Research, E. P. Xing, T. Jebara (eds.) (2014). (Vol. 32) 1188–1196.

  Many machine learning algorithms require the input to be represented as a fixed length feature vector. When it comes to texts, one of the most common representations is bag-of-words. Despite their popularity, bag-of-words models have two major weaknesses: they lose the ordering of the words and they also ignore semantics of the words. For example, "powerful," "strong" and "Paris" are equally distant. In this paper, we propose an unsupervised algorithm that learns vector representations of sentences and text documents. This algorithm represents each document by a dense vector which is trained to predict words in the document. Its construction gives our algorithm the potential to overcome the weaknesses of bag-of-words models. Empirical results show that our technique outperforms bag-of-words models as well as other techniques for text representations. Finally, we achieve new state-of-the-art results on several text classification and sentiment analysis tasks.
• Dropout: a simple way to prevent neural networks from overfitting. Srivastava, Nitish; Hinton, Geoffrey; Krizhevsky, Alex; Sutskever, Ilya; Salakhutdinov, Ruslan (2014). 15(1) 1929–1958.

  Deep neural nets with a large number of parameters are very powerful machine learning systems. However, overfitting is a serious problem in such networks. Large networks are also slow to use, making it difficult to deal with overfitting by combining the predictions of many different large neural nets at test time. Dropout is a technique for addressing this problem. The key idea is to randomly drop units (along with their connections) from the neural network during training. This prevents units from co-adapting too much. During training, dropout samples from an exponential number of different "thinned" networks. At test time, it is easy to approximate the effect of averaging the predictions of all these thinned networks by simply using a single unthinned network that has smaller weights. This significantly reduces overfitting and gives major improvements over other regularization methods. We show that dropout improves the performance of neural networks on supervised learning tasks in vision, speech recognition, document classification and computational biology, obtaining state-of-the-art results on many benchmark data sets.
• Generative Adversarial Networks. Goodfellow, Ian J.; Pouget-Abadie, Jean; Mirza, Mehdi; Xu, Bing; Warde-Farley, David; Ozair, Sherjil; Courville, Aaron; Bengio, Yoshua (2014).

  We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a mistake. This framework corresponds to a minimax two-player game. In the space of arbitrary functions G and D, a unique solution exists, with G recovering the training data distribution and D equal to 1/2 everywhere. In the case where G and D are defined by multilayer perceptrons, the entire system can be trained with backpropagation. There is no need for any Markov chains or unrolled approximate inference networks during either training or generation of samples. Experiments demonstrate the potential of the framework through qualitative and quantitative evaluation of the generated samples.
• Local Features in Image; Using HOG and SIFT. Shouno, Hayaru (2013). 67(3) 256–258.
• ランダムフォレスト. 波部斉 (2012). (Vol. 2012-CVIM-182)
• BIOMETRICS AUTHENTICATION TECHNIQUE FOR INTRUSION DETECTION SYSTEMS USING FINGERPRINT RECOGNITION. Mudholkar, Smita S.; Shende, Pradnya M.; Sarode, Milind V. (2012). 2(1) 57–65.

  Identifying attackers is a major apprehension to both organizations and governments. Recently, the most used applications for prevention or detection of attacks are intrusion detection systems. Biometrics technology is simply the measurement and use of the unique characteristics of living humans to distinguish them from one another and it is more useful as compare to passwords and tokens as they can be lost or stolen so we have choose the technique biometric authentication. The biometric authentication provides the ability to require more instances of authentication in such a quick and easy manner that users are not bothered by the additional requirements. In this paper, we have given a brief introduction about biometrics. Then we have given the information regarding the intrusion detection system and finally we have proposed a method which is based on fingerprint recognition which would allow us to detect more efficiently any abuse of the computer system that is running.
• MNIST handwritten digit database. LeCun, Yann; Cortes, Corinna (2010).
• Application of text mining(<Special feature>Data mining methodologies and application for libraries). YOSHIDA, Minoru; NAKAGAWA, Hiroshi (2010). 60(6) 230–235.
• Learning Multiple Layers of Features from Tiny Images Krizhevsky, Alex (2009). 32–33.
• Visualizing data using t-SNE. Van der Maaten, Laurens; Hinton, Geoffrey (2008). 9 2579–2605.

  We present a new technique called “t-SNE” that visualizes high-dimensional data by giving each datapoint a location in a two or three-dimensional map. The technique is a variation of Stochastic Neighbor Embedding (Hinton and Roweis, 2002) that is much easier to optimize, and produces significantly better visualizations by reducing the tendency to crowd points together in the center of the map. t-SNE is better than existing techniques at creating a single map that reveals structure at many different scales. This is particularly important for high-dimensional data that lie on several different, but related, low-dimensional manifolds, such as images of objects from multiple classes seen from multiple viewpoints. For visualizing the structure of very large data sets, we show how t-SNE can use random walks on neighborhood graphs to allow the implicit structure of all of the data to influence the way in which a subset of the data is displayed. We illustrate the performance of t-SNE on a wide variety of data sets and compare it with many other non-parametric visualization techniques, including Sammon mapping, Isomap, and Locally Linear Embedding. The visualiza- tions produced by t-SNE are significantly better than those produced by the other techniques on almost all of the data sets.
• The relationship between Precision-Recall and ROC curves. Davis, Jesse; Goadrich, Mark (2006). 233–240.
• Stochastic Neighbor Embedding. Hinton, Geoffrey E; Roweis, Sam S. Becker, S. Thrun, K. Obermayer (eds.) (2003). (Vol. 15) 857–864.
• 多変量解析に基づいた情報検索手法の比較検討. Technical Report (28(2001-FI-066), 竹内広宜; 小林メイ; 青野雅樹; 寒川光 (2002).
• Estimating the Support of a High-Dimensional Distribution. Scholkopf, Bernhard; Platt, John C.; Shawe-Taylor, John; Smola, Alex J.; Williamson, Robert C. (2001). 13(7) 1443–1471.

  Suppose you are given some data set drawn from an underlying probability distribution P and you want to estimate a "simple" subset S of input space such that the probability that a test point drawn from P lies outside of S equals some a priori specified value between 0 and 1. We propose a method to approach this problem by trying to estimate a function f that is positive on S and negative on the complement. The functional form of f is given by a kernel expansion in terms of a potentially small subset of the training data; it is regularized by controlling the length of the weight vector in an associated feature space. The expansion coefficients are found by solving a quadratic programming problem, which we do by carrying out sequential optimization over pairs of input patterns. We also provide a theoretical analysis of the statistical performance of our algorithm. The algorithm is a natural extension of the support vector algorithm to the case of unlabeled data.
• The information bottleneck method. Tishby, Naftali; Pereira, Fernando C.; Bialek, William (2000).

  We define the relevant information in a signal \($x\in X$\) as being the information that this signal provides about another signal \($y\in \Y$\). Examples include the information that face images provide about the names of the people portrayed, or the information that speech sounds provide about the words spoken. Understanding the signal \($x$\) requires more than just predicting \($y$\), it also requires specifying which features of \($\X$\) play a role in the prediction. We formalize this problem as that of finding a short code for \($\X$\) that preserves the maximum information about \($\Y$\). That is, we squeeze the information that \($\X$\) provides about \($\Y$\) through a `bottleneck' formed by a limited set of codewords \($\tX$\). This constrained optimization problem can be seen as a generalization of rate distortion theory in which the distortion measure \($d(x,\x)$\) emerges from the joint statistics of \($\X$\) and \($\Y$\). This approach yields an exact set of self consistent equations for the coding rules \($X \to \tX$\) and \($\tX \to \Y$\). Solutions to these equations can be found by a convergent re-estimation method that generalizes the Blahut-Arimoto algorithm. Our variational principle provides a surprisingly rich framework for discussing a variety of problems in signal processing and learning, as will be described in detail elsewhere.
• Globally convergent block-coordinate techniques for unconstrained optimization. Grippof, Luigi; Sciandrone, M. (1999). 10 587–637.
• Backpropagation through time: what it does and how to do it. Werbos, P. J. (1990). 78(10) 1550–1560.