Recognizing places from an opposing viewpoint during a return trip is a common experience for human drivers. However, the analogous robotics capability, visual place recognition (VPR) with limited field of view cameras under 180 degree rotations, has proven to be challenging to achieve. To address this problem, this paper presents Same Place Opposing Trajectory (SPOT), a technique for opposing viewpoint VPR that relies exclusively on structure estimated through stereo visual odometry (VO). The method extends recent advances in lidar descriptors and utilizes a novel double (similar and opposing) distance matrix sequence matching method. We evaluate SPOT on a publicly available dataset with 6.7-7.6 km routes driven in similar and opposing directions under various lighting conditions. The proposed algorithm demonstrates remarkable improvement over the state-of-the-art, achieving up to 91.7% recall at 100% precision in opposing viewpoint cases, while requiring less storage than all baselines tested and running faster than all but one. Moreover, the proposed method assumes no a priori knowledge of whether the viewpoint is similar or opposing, and also demonstrates competitive performance in similar viewpoint cases.

The RecSys 2023 Challenge involved a conversion prediction task in the online advertising space. The dataset was provided by ShareChat (Mohalla Tech Pvt Ltd). The challenge data represents a sample of ad impressions served to the users over a period of 22 days and the task is for a given ad impression, to predict a conversion (install an app) will happen or not. The challenge ran for 3 months with a public dashboard. There were 519 teams registered and 231 teams made at least one submission. The task setting represents an important research area of modeling ad recommendations under user privacy. We identify interesting themes in feature engineering, addressing sparsity and calibrating across multi-step predictions.

Yaobo Liang, Nan Duan, Yeyun Gong, Ning Wu, Fenfei Guo, Weizhen Qi, Ming Gong, Linjun Shou, Daxin Jiang, Guihong Cao, Xiaodong Fan, Ruofei Zhang, Rahul Agrawal, Edward Cui, Sining Wei, Taroon Bharti, Ying Qiao, Jiun-Hung Chen, Winnie Wu, Shuguang Liu, Fan Yang, Daniel Campos, Rangan Majumder, Ming Zhou. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). 2020.

Unsupervised multi-lingual language modeling has gained attraction in the last few years and poly-lingual topic models provide a mechanism to learn aligned document representations. However, training such models require translation-aligned data across languages, which is not always available. Also, in case of short texts like tweets, search queries, etc, the training of topic models continues to be a challenge. In this work, we present a novel strategy of creating a pseudo-parallel dataset followed by training topic models for sponsored search retrieval, that also mitigates the short text challenge. Our data augmentation strategy leverages easily available bipartite click-though graph that allows us to draw similar documents in different languages. The proposed methodology is evaluated on sponsored search system whose performance is measured on correctly matching the user intent, presented via the query, with ads provided by the advertiser. Our experiments substantiate the goodness of the method on EuroParl dataset and live search-engine traffic.

The measure of similarity between nodes in a graph is a useful tool in many areas of computer science. SimRank, proposed by Jeh and Widom [7], is a classic measure of similarities of nodes in graph that has both theoretical and intuitive properties and has been extensively studied and used in many applications such as Query-Rewriting, link prediction, collaborative filtering and so on. Existing works based on Simrank primarily focus on preserving the microscopic structure, such as the second and third order proximity of the vertices, while the macroscopic scale-free property is largely ignored. Scale-free property is a critical property of any real-world web graphs where the vertex degrees follow a heavy-tailed distribution. In this paper, we introduce P-Simrank which extends the idea of Simrank to Scale-free bipartite networks. To study the efficacy of the proposed solution on a real world problem, we tested the same on the well known query-rewriting problem in sponsored search domain using bipartite click graph, similar to Simrank++ [1], which acts as our baseline. We show that Simrank++ produces sub-optimal similarity scores in case of bipartite graphs where degree distribution of vertices follow power-law. We also show how P-Simrank can be optimized for real-world large graphs. Finally, we experimentally evaluate P-Simrank algorithm against Simrank++, using actual click graphs obtained from Bing, and show that P-Simrank outperforms Simrank++ in variety of metrics.

The two major roadblocks for state of the art Internet of Things (IoT) infrastructure like smart buildings, smart cities, etc. are lack of trust between various entities of system and single point of failure which is a vulnerability causing extreme damage to the whole system. This paper proposes a blockchain based IoT security solution where, trust is established through the immutable and decentralized nature of blockchain. The distributed nature of blockchain makes the system more robust and immune to single point of failure. We propose a mechanism to establish continuous security in the system by evaluating legitimate presence of user in valid IoT-Zone continuously without user intervention. Every user interaction in an IoT environment is stored in blockchain as a transaction and series of these transactions represent a user's IoT-trail. A unique digital crypto-token is required for a user interaction to be legitimate. This token is used as an access control mechanism to prevent any unauthorized access to the system. Tokens are pre-generated using a prediction model based on user's IoT-trail in the blockchain. By using blockchain as an underlying framework in IoT environment and through the method of continuous security, we made the system more secure, robust and interoperable.

The objective in extreme multi-label learning is to build classifiers that can annotate a data point with the subset of relevant labels from an extremely large label set. Extreme classification has, thus far, only been studied in the context of predicting labels for novel test points. This paper formulates the extreme classification problem when predictions need to be made on training points with partially revealed labels. This allows the reformulation of warm-start tagging, ranking and recommendation problems as extreme multi-label learning with each item to be ranked/recommended being mapped onto a separate label. The SwiftXML algorithm is developed to tackle such warm-start applications by leveraging label features. SwiftXML improves upon the state-of-the-art tree based extreme classifiers by partitioning tree nodes using two hyperplanes learnt jointly in the label and data point feature spaces. Optimization is carried out via an alternating minimization algorithm allowing SwiftXML to efficiently scale to large problems.

This paper develops the Parabel algorithm for extreme multi-label learning where the objective is to learn classifiers that can annotate each data point with the most relevant subset of labels from an extremely large label set. The state-of-the-art 1-vs-All based DiSMEC and PPDSparse algorithms are the most accurate but can take upto months for training and prediction as they learn and apply an independent linear classifier per label. Consequently, they do not scale to large datasets with millions of labels. Parabel addresses both limitations by learning a balanced label hierarchy such that: (a) the 1-vs-All classifiers in the leaf nodes of the label hierarchy can be trained on a small subset of the training set thereby reducing the training time to a few hours on a single core of a standard desktop and (b) novel points can be classified by traversing the learned hierarchy in logarithmic time and applying the 1-vs-All classifiers present in just the leaf thereby reducing the prediction time to a few milliseconds per test point. This allows Parabel to scale to tasks considered infeasible for DiSMEC and PPDSparse such as predicting the subset of 7 million Bing queries that might lead to a click on a given ad-landing page for dynamic search advertising. Experiments on multiple benchmark datasets revealed that Parabel could be almost as accurate as PPDSparse and DiSMEC while being upto 1,000x faster at training and upto 40x-10,000x faster at prediction. Furthermore, Parabel was demonstrated to significantly improve dynamic search advertising on Bing by more than doubling the ad recall and improving the click-through rate by 20%. Source code for Parabel can be downloaded from [1].

Determining reputation of an advertiser in sponsored search is a recent important problem with direct impact on revenue for web publishers and relevance of ads. Individual performance of advertisers is usually expressed through observed click through rate, which depends on advertiser reputation, ad relevance and position. However, advertiser reputation has not been explicitly modeled in click prediction literature. Using traditional approaches in web page popularity for organic search in this context is not reasonable as the notion of link-structure in web is not directly applicable to sponsored search. In this study, we motivate and propose a pairwise preference relation model to study the advertiser reputation problem. Pairwise comparisons of advertisers give information over and above the information available in their individual historical performances. We relate the notion of preference among the advertisers to the spectral properties of the preference graph. We provide empirical evidence of the existence of reputation bias in click behavior. Consequently, we experiment with this signal to improve click prediction.

Recommending phrases from web pages for advertisers to bid on against search engine queries is an important research problem with direct commercial impact. Most approaches have found it infeasible to determine the relevance of all possible queries to a given ad landing page and have focussed on making recommendations from a small set of phrases extracted (and expanded) from the page using NLP and ranking based techniques. In this paper, we eschew this paradigm, and demonstrate that it is possible to efficiently predict the relevant subset of queries from a large set of monetizable ones by posing the problem as a multi-label learning task with each query being represented by a separate label.

We consider the problem of estimating rates of rare events for high dimensional, multivariate categorical data where several dimensions are hierarchical. Such problems are routine in several data mining applications including computational advertising, our main focus in this paper. We propose LMMH, a novel log-linear modeling method that scales to massive data applications with billions of training records and several million potential predictors in a map-reduce framework. Our method exploits correlations in aggregates observed at multiple resolutions when working with multiple hierarchies; stable estimates at coarser resolution provide informative prior information to improve estimates at finer resolutions. Other than prediction accuracy and scalability, our method has an inbuilt variable screening procedure based on a "spike and slab prior" that provides parsimony by removing non-informative predictors without hurting predictive accuracy. We perform large scale experiments on data from real computational advertising applications and illustrate our approach on datasets with several billion records and hundreds of millions of predictors. Extensive comparisons with other benchmark methods show significant improvements in prediction accuracy.