In the first part of the talk, I will present a fully statistical, data-driven, instance-based approach to question answering (IBQA) that learns how to answer new questions from similar training questions and their known correct answers. We represent training questions as points in a multi-dimensional space and cluster them according to different granularity, scatter, and similarity metrics. From each individual cluster we automatically learn an answering strategy for finding answers to questions. When answering a new question that is covered by several clusters, multiple answering strategies are simultaneously employed. The resulting answer confidence combines elements such as each strategy's estimated probability of success, cluster similarity to the new question, cluster size, and cluster granularity. The IBQA approach obtains good performance on factoid and definitional questions, comparable to the performance of top systems participating in official question answering evaluations. The second part of the presentation will be an overview of natural language processing and machine learning projects going on in our group at Siemens Medical Solutions. We perform application driven research, seeking statistical and machine learning solutions for different clinical problems, and targeting real products. In natural language processing, we work on information extraction and domain-specific question answering from medical text (patient records, medical knowledge), integrating unstructured and semi-structured sources. Another area of interest is large-scale text classification for medical coding, which several existing solutions approach from a rule-based perspective. Another set of projects in our group focuses on computer-aided diagnosis using computer vision and classification based on data sources such as fMri, ultrasound, and x-ray data for detection and classification for disease diagnosis (e.g. lung cancer and colon cancer).