Learning from Multiple Outlooks
نویسندگان
چکیده
We propose a novel problem formulation of learning a single task when the data are provided in different feature spaces. Each such space is called an outlook, and is assumed to contain both labeled and unlabeled data. The objective is to take advantage of the data from all the outlooks to better classify each of the outlooks. We devise an algorithm that computes optimal affine mappings from different outlooks to a target outlook by matching moments of the empirical distributions. We further derive a probabilistic interpretation of the resulting algorithm and a sample complexity bound indicating how many samples are needed to adequately find the mapping. We report the results of extensive experiments on activity recognition tasks that show the value of the proposed approach in boosting performance.
منابع مشابه
Learning-Based Assume-Guarantee Regression Verification
Due to enormous resource consumption, model checking each revision of evolving systems repeatedly is impractical. To reduce cost in checking every revision, contextual assumptions are reused from assumeguarantee reasoning. However, contextual assumptions are not always reusable. We propose a fine-grained learning technique to maximize the reuse of contextual assumptions. Based on fine-grained l...
متن کاملReaching Today's Information Security Students
Classes at university today comprise students from the Baby Boomers, Generation X and Y. The different outlooks on life of these generations affect their choice of education options and their learning preferences. There are numerous ways academics can innovatively deliver Information Security learning materials that meet the needs of these generations, whilst still achieving the educational goa...
متن کاملThe Predictability of Cane Production in the South African Sugar Industry Using Seasonal Climate Outlooks and the Canesim Yield Forecasting System
Timely and accurate yield forecasts prior to and during the milling season present opportunities to improve various industry activities, such as milling operations, international trade and agronomic optimisation. The Canesim model-based yield forecasting system was used to quantify prediction skills at different times of the year, using historic climate data and a history of seasonal climate ou...
متن کاملSub-Space Clustering and Evidence Accumulation for Unsupervised Network Anomaly Detection
Network anomaly detection has been a hot research topic for many years. Most detection systems proposed so far employ a supervised strategy to accomplish the task, using either signature-based detection methods or supervised-learning techniques. However, both approaches present major limitations: the former fails to detect unknown anomalies, the latter requires training and labeled traffic, whi...
متن کاملRating Transitions and Defaults Conditional on Watchlist, Outlook and Rating History
This report documents global corporate credit rating transition and default rates during the 1995-2003 period conditional on the full information in Moody's published credit opinions, which, in addition to the current credit rating, include prior rating actions and current rating outlooks and reviews. The primary findings of this study are: • Moody's rating system management practices attempt t...
متن کامل