Software systems that learn from user data with machine learning (ML) have become ubiquitous over the last years. Recent law such as the "General Data Protection Regulation" (GDPR) requires organisations that process personal data to delete user data upon request (enacting the "right to be forgotten"). However, this regulation does not only require the deletion of user data from databases, but also applies to ML models that have been learned from the stored data. We therefore argue that ML applications should offer users to unlearn their data from trained models in a timely manner. We explore how fast this unlearning can be done under the constraints imposed by real world deployments, and introduce the problem of low-latency machine unlearning: maintaining a deployed ML model in-place under the removal of a small fraction of training samples without retraining.

Apache Mahout is a library for scalable machine learning (ML) on distributed dataflow systems, offering various implementations of classification, clustering, dimensionality reduction and recommendation algorithms. Mahout was a pioneer in large-scale machine learning in 2008, when it started and targeted MapReduce, which was the predominant abstraction for scalable computing in industry at that time. Mahout has been widely used by leading web companies and is part of several commercial cloud offerings. In recent years, Mahout migrated to a general framework enabling a mix of dataflow programming and linear algebraic computations on backends such as Apache Spark and Apache Flink. This design allows users to execute data preprocessing and model training in a single, unified dataflow system, instead of requiring a complex integration of several specialized systems. Mahout is maintained as a community-driven open source project at the Apache Software Foundation, and is available under https://mahout.apache.org.

Recommender systems are ubiquitous in the modern internet, where they help users find items they might like. A widely deployed recommendation approach is item-based collaborative filtering. This approach relies on analyzing large item cooccurrence matrices that denote how many users interacted with a pair of items. The potentially quadratic number of items to compare poses a scalability bottleneck in analyzing such item cooccurrences. Additionally, this problem intensifies in real world use cases with incrementally growing datasets, especially when the recommendation model is regularly recomputed from scratch. We highlight the connection between the growing cost of item-based recommendation and densification processes in common interaction datasets. Based on our findings, we propose an efficient incremental algorithm for item-based collaborative filtering based on cooccurrence analysis. This approach restricts the number of interactions to consider from 'power users' and 'ubiquitous items' to guarantee a provably constant amount of work per user-item interaction to process. We discuss efficient implementations of our algorithm on a single machine as well as on a distributed stream processing engine, and present an extensive experimental evaluation. Our results confirm the asymptotic benefits of the incremental approach. Furthermore, we find that our implementation is an order of magnitude faster than existing open source recommender libraries on many datasets, and at the same time scales to high dimensional datasets which these existing recommenders fail to process.

We feel strongly that a contemporary introductory course in machine organization and assembly language should focus on the essentials of how computers execute programs, and not be distracted by the complications of the extraordinarily sophisticated microprocessors that are available today. These essentials should form a strong base of knowledge from which students can draw as they continue their education in computer science. Ideally these goals should be attained in an environment that fosters experimentation and cooperation, and with the aid of projects that generate interest and enthusiasm among the students. We have developed and are currently teaching a course at New Mexico State University that meets many of these goals. The course concentrates on a simple but relatively complete microprocessor architecture, that of the Motorola 68HC11 processor. Three different teaching techniques are used to encourage experimentation and team work: learning sessions, simulator labs, and microprocessor labs. New concepts are introduced in learning sessions, which combine traditional lecturing with student exploration. The understanding of these new concepts is strengthened through labs and assignments. Simulator labs and assignments, which require interaction with a simulator of the Motorola 68HC11 microprocessor, focus on the 68HC11's instruction set architecture. Microprocessor labs and assignments, which essentially are designing and building sessions, focus on the use of a 68HC11 microprocessor to control a motorized vehicle. During microprocessor labs students populate printed circuit cards, build motorized vehicles (or other roboticized exotica), and design and implement assembly language programs that provide communication between a personal computer and a 68HC11 processor, and a 68HC11 processor and a motorized vehicle. We have found that the costs of running this course are minimal and the results are very favorable in terms of student enthusiasm and achievement.