Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

This tutorial is concerned with a comprehensive introduction to the research area of learning to rank for information retrieval. In the first part of the tutorial, we will introduce three major approaches to learning to rank, i.e., the pointwise, pairwise, and listwise approaches, analyze the relationship between the loss functions used in these approaches and the widely-used IR evaluation measures, evaluate the performance of these approaches on the LETOR benchmark datasets, and demonstrate how to use these approaches to solve real ranking applications. In the second part of the tutorial, we will discuss some advanced topics regarding learning to rank, such as relational ranking, diverse ranking, semi-supervised ranking, transfer ranking, query-dependent ranking, and training data preprocessing. In the third part, we will briefly mention the recent advances on statistical learning theory for ranking, which explain the generalization ability and statistical consistency of different ranking methods. In the last part, we will conclude the tutorial and show several future research directions.

/pdf/learning-to-rank-for-information-retrieval-2q6t3viyug.pdf

Learning to Rank for Information Retrieval

Urbanization's rapid progress has modernized many people's lives but also engendered big issues, such as traffic congestion, energy consumption, and pollution. Urban computing aims to tackle these issues by using the data that has been generated in cities (e.g., traffic flow, human mobility, and geographical data). Urban computing connects urban sensing, data management, data analytics, and service providing into a recurrent process for an unobtrusive and continuous improvement of people's lives, city operation systems, and the environment. Urban computing is an interdisciplinary field where computer sciences meet conventional city-related fields, like transportation, civil engineering, environment, economy, ecology, and sociology in the context of urban spaces. This article first introduces the concept of urban computing, discussing its general framework and key challenges from the perspective of computer sciences. Second, we classify the applications of urban computing into seven categories, consisting of urban planning, transportation, the environment, energy, social, economy, and public safety and security, presenting representative scenarios in each category. Third, we summarize the typical technologies that are needed in urban computing into four folds, which are about urban sensing, urban data management, knowledge fusion across heterogeneous data, and urban data visualization. Finally, we give an outlook on the future of urban computing, suggesting a few research topics that are somehow missing in the community.

/pdf/urban-computing-concepts-methodologies-and-applications-1wix0tsb38.pdf

Urban Computing: Concepts, Methodologies, and Applications

GPS-equipped taxis can be regarded as mobile sensors probing traffic flows on road surfaces, and taxi drivers are usually experienced in finding the fastest (quickest) route to a destination based on their knowledge. In this paper, we mine smart driving directions from the historical GPS trajectories of a large number of taxis, and provide a user with the practically fastest route to a given destination at a given departure time. In our approach, we propose a time-dependent landmark graph, where a node (landmark) is a road segment frequently traversed by taxis, to model the intelligence of taxi drivers and the properties of dynamic road networks. Then, a Variance-Entropy-Based Clustering approach is devised to estimate the distribution of travel time between two landmarks in different time slots. Based on this graph, we design a two-stage routing algorithm to compute the practically fastest route. We build our system based on a real-world trajectory dataset generated by over 33,000 taxis in a period of 3 months, and evaluate the system by conducting both synthetic experiments and in-the-field evaluations. As a result, 60-70% of the routes suggested by our method are faster than the competing methods, and 20% of the routes share the same results. On average, 50% of our routes are at least 20% faster than the competing approaches.

/pdf/t-drive-driving-directions-based-on-taxi-trajectories-1xb3zuw0ip.pdf

T-drive: driving directions based on taxi trajectories

The Concept Of Law

We consider the problem of large scale retrieval evaluation. Recently two methods based on random sampling were proposed as a solution to the extensive effort required to judge tens of thousands of documents. While the first method proposed by Aslam et al. [1] is quite accurate and efficient, it is overly complex, making it difficult to be used by the community, and while the second method proposed by Yilmaz et al., infAP [14], is relatively simple, it is less efficient than the former since it employs uniform random sampling from the set of complete judgments. Further, none of these methods provide confidence intervals on the estimated values.The contribution of this paper is threefold: (1) we derive confidence intervals for infAP, (2) we extend infAP to incorporate nonrandom relevance judgments by employing stratified random sampling, hence combining the efficiency of stratification with the simplicity of random sampling, (3) we describe how this approach can be utilized to estimate nDCG from incomplete judgments. We validate the proposed methods using TREC data and demonstrate that these new methods can be used to incorporate nonrandom samples, as were available in TREC Terabyte track '06.

/pdf/a-simple-and-efficient-sampling-method-for-estimating-ap-and-4z9q3o12st.pdf

A simple and efficient sampling method for estimating AP and NDCG

This paper proposes and solves the Time-Interval All Fastest Path (allFP) query. Given a user-defined leaving or arrival time interval I, a source node s and an end node e, allFP asks for a set of all fastest paths from s to e, one for each sub-interval of I. Note that the query algorithm should find a partitioning of I into sub-intervals. Existing methods can only be used to solve a very special case of the problem, when the leaving time is a single time instant. A straightforward solution to the allFP query is to run existing methods many times, once for every time instant in I. This paper proposes a solution based on novel extensions to the A* algorithm. Instead of expanding the network many times, we expand once. The travel time on a path is kept as a function of leaving time. Methods to combine travel-time functions are provided to expand a path. A novel lower-bound estimator for travel time is proposed. Performance results reveal that our method is more efficient and more accurate than the discrete-time approach.

/pdf/finding-fastest-paths-on-a-road-network-with-speed-patterns-3i1qpen9e4.pdf

Finding Fastest Paths on A Road Network with Speed Patterns

This paper presents results comparing user preference for search engine rankings with measures of effectiveness computed from a test collection. It establishes that preferences and evaluation measures correlate: systems measured as better on a test collection are preferred by users. This correlation is established for both "conventional web retrieval" and for retrieval that emphasizes diverse results. The nDCG measure is found to correlate best with user preferences compared to a selection of other well known measures. Unlike previous studies in this area, this examination involved a large population of users, gathered through crowd sourcing, exposed to a wide range of retrieval systems, test collections and search tasks. Reasons for user preferences were also gathered and analyzed. The work revealed a number of new results, but also showed that there is much scope for future work refining effectiveness measures to better capture user preferences.

/pdf/do-user-preferences-and-evaluation-measures-line-up-2navypwhdq.pdf

Do user preferences and evaluation measures line up

: The Million Query (1MQ) track ran for the second time in TREC 2008. The track is designed to serve two purposes: first, it is an exploration of ad-hoc retrieval over a large set of queries and a large collection of documents; second, it investigates questions of system evaluation, in particular whether it is better to evaluate using many shallow judgments or fewer thorough judgments. As with the 2007 track [ACA+07], participants ran 10,000 queries against a collection of 25 million documents. The 2008 track differed in the following ways: 1. Queries were assigned to one of four categories. 2. Each query was assigned a target of 8, 16, 32, 64, or 128 judgments. 3. Assessors could judge documents \not relevant but reasonable". Section 1 describes how the corpus and queries were selected, the query classes, details of the submission formats, and a brief description of each submitted run. Section 2 provides an overview of the judging process, including a sketch of how it alternated between two methods for selecting the small set of documents to be judged. Sections 3.1 and 3.2 provide an overview of those two selection methods, developed at UMass and NEU, respectively. In Section 4 we present statistics collected during the judging process, including the total number of queries judged, how many judgments were served by each approach, and so on, along with the overall results of the track. We present additional results and analysis in Section 5.

/pdf/million-query-track-2008-overview-5hdy8vk3yg.pdf

Million Query Track 2008 Overview

Given a set O of weighted objects, a set S of sites, and a query site s, the bichromatic RNN query computes the influence set of s, or the set of objects in O that consider s as the nearest site among all sites in S. The influence of a site s can be defined as the total weight of its RNNs. This paper addresses the new and interesting problem of finding the top-t most influential sites from S, inside a given spatial region Q. A straightforward approach is to find the sites in Q, and compute the RNNs of every such site. This approach is not efficient for two reasons. First, all sites in Q need to be identified whatsoever, and the number may be large. Second, both the site R-tree and the object R-tree need to be browsed a large number of times. For each site in Q, the R-tree of sites is browsed to identify the influence region -- a polygonal region that may contain RNNs, and then the R-tree of objects is browsed to find the RNN set. This paper proposes an algorithm called TopInfluential-Sites, which finds the top-t most influential sites by browsing both trees once systematically. Novel pruning techniques are provided, based on a new metric called minExistDNN. There is no need to compute the influence for all sites in Q, or even to visit all sites in Q. Experimental results verify that our proposed method outperforms the straightforward approach.

/pdf/on-computing-top-t-most-influential-spatial-sites-2bxglhsmr6.pdf

Evangelos Kanoulas

Papers

A simple and efficient sampling method for estimating AP and NDCG

Finding Fastest Paths on A Road Network with Speed Patterns

Do user preferences and evaluation measures line up

Million Query Track 2008 Overview

On computing top- t most influential spatial sites