This paper presents ASGARD, a novel generative framework for personalized advertising strategy design. ASGARD integrates user-preferences while maintaining the ability to autonomously generate high-quality outputs. The core challenge is balancing user preferences with the model's knowledge, enabling it to align with user inputs or propose optimal alternatives. Our novel approach uses a token-driven method for strategy generation, adapts a token masking strategy for training, and refines the loss function to prevent issues like mode collapse. We evaluate ASGARD to demonstrate its effectiveness, identify limitations, and suggest future enhancements. To our knowledge, this is the first approach that meets all constraints of advertising strategy design while allowing user-preference integration, showing potential for generalization across transformer-based generative models.

Digital advertising has become one of the most important aspects of modern marketing, as it allows businesses to reach larger audiences with greater precision, controllable cost, and measurable feedback. However, the process of designing effective advertising strategies relies heavily on human experts and thus remains suboptimal. In this paper, we propose a novel method for Contextual Advertising Strategy Generation via Attention and InteRaction Guidance (ASGAR) that leverages transformers and a soft contrastive learning approach to optimize campaign performance. An advertising strategy is a combination of multiple targeting options, and its performance is tied strictly to the combination as a whole. This makes the exploration of the high combinatorial space infeasible and autoregressive methods inefficient. Therefore, constraints of non-combinatorial exploration and non-autoregressive generation have to be met. To the best of our knowledge, this is the first method that satisfies all constraints while also outperforming the previous methods. We compare our results with state-of-the-art methods on a public data set and with human experts in a company-deployed environment. We show that our method can effectively generate high-performance advertising strategies with better stability and controllable exploration.

This paper improves the conventional sparse representation based classification (SRC) method, through incorporating wavelet coefficients. For this reason, the proposed method is called Sparse Representation Wavelet based Classification (SRWC). In the present study, we fuse the image features described by the complementary information from the low sub-band of the wavelet coefficients and sparse representation to outperform the conventional SRC according to accuracy. This holds because the wavelets promote sparsity and provide structural information about the image, which increases the accuracy of classification. To validate the capabilities and underline the advantages of the novel SRWC, we conducted an extensive number of experiments using publicly available datasets and compared our results with contemporary methods.

Nowadays, in the Web 2.0 reality, one of the most challenging task for companies that aim to manage and recommend job offers is to convey this enormous amount of information in a succinct and intelligent manner such to increase the performances of matching operations against users profiles/curricula and optimize the time/space complexity of these processes. With this goal, this paper presents a novel method to formalize the textual content of job offers that aims at identifying the most relevant information and fields expressed by them and leverage this compact formalization for job recommendation and profile matching in social network environments. This method has been then developed and tested in the industrial environment represented by Multiposting and Work4, world leaders in digital solutions of e-recruitment problems. In this study three classes of documents are considered: job offers, job categories and social network user profiles (as potential job candidates); each class contains several fields with textual information. The proposed representation method permits to dynamically identify those text fields, for each class, that could help a cross-matching strategy in order to preserve, from one hand, the matching/recommendation performances and, on the other hand, reduce the cost of these operations (due to a straightforward dimensionality reduction mechanism). We then evaluated and compared the presented approach showing significant improvements on both categorization and recommendation tasks by also drastically reducing their computational costs.

This paper presents taxonomy-based recommender systems that propose relevant jobs to Facebook and LinkedIn users; they are being developed by Work4, a San Francisco-based software company and the Global Leader in Social and Mobile Recruiting that offers Facebook recruitment solutions; to use its applications, Facebook or LinkedIn users explicitly grant access to some parts of their data, and they are presented with the jobs whose descriptions are matching their profiles the most. In this paper, we use the O*NET-SOC taxonomy, a taxonomy that defines the set of occupations across the world of work, to develop a new taxonomy-based vector model for social network users and job descriptions suited to the task of job recommendation; we propose two similarity functions based on the AND and OR fuzzy logic's operators, suited to the proposed vector model. We compare the performance of our proposed vector model to the TF-IDF model using our proposed similarity functions and the classic heuristic measures; the results show that the taxonomy-based vector model outperforms the TF-IDF model. We then use SVMs (Support Vector Machines) with a mechanism to handle unbalanced datasets, to learn similarity functions from our data; the learnt models yield better results than heuristic similarity measures. The comparison of our methods to two methods of the literature (a matrix factorization method and the Collaborative Topic Regression) shows that our best method yields better results than those two methods in terms of AUC. The proposed taxonomy-based vector model leads to an efficient dimensionality reduction method in the task of job recommendation.

This paper presents a content-based recommender system which proposes jobs to Facebook and LinkedIn users. A variant of this recommender system is currently used by Work4, a San Francisco-based software company that offers Facebook recruitment solutions.

Prediction of user behavior in Social Networks is important for a lot of applications, ranging from marketing to social community management. In this paper, we develop and test a model to estimate the propensity of a user to stop using the social platform in a near future. This problem is called churn prediction and has been extensively studied in telecommunication networks. We focus here on building a statistical model estimating the probability that a user will leave the social network in the near future. The model is based on graph attributes extracted in the user's vicinity. We present a novel algorithm to accurately detect overlapping local communities in social graphs. Our algorithm outperforms the state of the art methods and is able to deal with pathological cases which can occur in real networks. We show that using attributes computed from the local community around the user allows to build a robust statistical model to predict churn. Our ideas are tested on one of the largest French social blog platform, Sky rock, where millions of teenagers interact daily.

Natural Language Processing has emerged as an active field of research in the machine learning community. Several methods based on statistical information have been proposed. However, with the linguistic complexity of the texts, semantic-based approaches have been investigated. In this paper, we propose a Semantic Kernel for semi- structured biomedical documents. The semantic meanings of words are extracted using the UMLS framework. The kernel, with a SVM classifier, has been applied to a text categorization task on a medical corpus of free text documents. The results have shown that the Semantic Kernel outperforms the Linear Kernel and the Naive Bayes classifier. Moreover, this kernel was ranked in the top ten of the best algorithms among 44 classification methods at the 2007 CMC Medical NLP International Challenge.

Sequential pattern mining allows to discover temporal relationship between items within a database. The patterns can then be used to generate association rules. When the databases are very large, the execution speed and the memory usage of the mining algorithm become critical parameters. Previous research has focused on either one of the two parameters. In this paper, we present bitSPADE, a novel algorithm that combines the best features of SPAM, one of the fastest algorithm, and SPADE, one of the most memory efficient algorithm. Moreover, we introduce a new pruning strategy that enables bitSPADE to reach high performances. Experimental evaluations showed that bitSPADE ensures an efficient tradeoff between speed and memory usage by outperforming SPADE by both speed and memory usage factors more than 3.4 and SPAM by a memory consumption factor up to more than an order of magnitude.

In practical applications, recognition accuracy is sometimes not the only criterion; capability to reject erroneous patterns might also be needed. We show that there is a trade-off between these two properties. An efficient solution to this trade-off is brought about by the use of different algorithms implemented in various modules, i.e. multi-modular architectures. We present a general mechanism for designing and training multi-modular architectures, integrating various neural networks into a unique pattern recognition system, which is globally trained. It is possible to realize, within the system, feature extraction and recognition in successive modules which are cooperatively trained. We discuss various rejection criteria for neural networks and multi-modular architectures. We then give two examples of such systems, study their rejection capabilities and show how to use them for segmentation. In handwritten optical character recognition, our system achieves performances at state-of-the-art level, but is eight times faster. In human face recognition, our system is intended to work in the real world.