Graph Neural Networks (GNNs) achieve high performance in various real-world applications, such as drug discovery, traffic states prediction, and recommendation systems. The fact that building powerful GNNs requires a large amount of training data, powerful computing resources, and human expertise turns the models into lucrative targets for model stealing attacks. Prior work has revealed that the threat vector of stealing attacks against GNNs is large and diverse, as an attacker can leverage various heterogeneous signals ranging from node labels to high-dimensional node embeddings to create a local copy of the target GNN at a fraction of the original training costs. This diversity in the threat vector renders the design of effective and general defenses challenging and existing defenses usually focus on one particular stealing setup. Additionally, they solely provide means to identify stolen model copies rather than preventing the attack. To close this gap, we propose the first and general Active Defense Against GNN Extraction (ADAGE). By analyzing the queries to the GNN, tracking their diversity in terms of proximity to different communities identified in the underlying graph, and increasing the defense strength with the growing fraction of communities that have been queried, ADAGE can prevent stealing in all common attack setups. Our extensive experimental evaluation using six benchmark datasets, four GNN models, and three types of adaptive attackers shows that ADAGE penalizes attackers to the degree of rendering stealing impossible, whilst not harming predictive performance for legitimate users. ADAGE, thereby, contributes towards securely sharing valuable GNNs in the future.

Agent-based models (ABMs) are valuable for modelling complex, potentially out-of-equilibria scenarios. However, ABMs have long suffered from the Lucas critique, stating that agent behaviour should adapt to environmental changes. Furthermore, the environment itself often adapts to these behavioural changes, creating a complex bi-level adaptation problem. Recent progress integrating multi-agent reinforcement learning into ABMs introduces adaptive agent behaviour, beginning to address the first part of this critique, however, the approaches are still relatively ad hoc, lacking a general formulation, and furthermore, do not tackle the second aspect of simultaneously adapting environmental level characteristics in addition to the agent behaviours. In this work, we develop a generic two-layer framework for ADaptive AGEnt based modelling (ADAGE) for addressing these problems. This framework formalises the bi-level problem as a Stackelberg game with conditional behavioural policies, providing a consolidated framework for adaptive agent-based modelling based on solving a coupled set of non-linear equations. We demonstrate how this generic approach encapsulates several common (previously viewed as distinct) ABM tasks, such as policy design, calibration, scenario generation, and robust behavioural learning under one unified framework. We provide example simulations on multiple complex economic and financial environments, showing the strength of the novel framework under these canonical settings, addressing long-standing critiques of traditional ABMs.

“It costs five times more to acquire a customer than to retain a customer” – goes a famous adage. With due respect to this adage, if you don’t acquire new customers, your bank would cease to grow. So, one cannot overemphasize the importance of customer acquisition. Today Artificial Intelligence has permeated almost every aspect of the banking and financial industry. It looks like customer acquisition just got the Conversational AI reinforcement. The banking customer acquisition process slowly shifted gears from physical to digital channels in the last decade. Today, banking and financial players can speed up the customer acquisition cycle and reduce the costs associated, thanks to Conversational AI. But legacy banks need to move fast as there is a constant threat of upstart fintech players giving banks a run for their money! Meet your customers where they are The distribution-led ‘bank branch’ growth model has few takers today. In the first decade of this millennium, the number of physical branches closely linked to banking customer base and revenues. Not anymore. Today’s digital-first customers expect you to meet them where they are – web, social, instant messenger, apps, etc. Conversational AI can help you be omnipresent on digital channels to capture your customer’s mindshare and, ultimately, the wallet share in its many avatars. Reduce the friction in customer journeys Every prospective banking customer would go through a handful of predefined digital customer journeys. It could start searching on Google, comparing competitor offerings, chatting with reps online, checking for social proof, etc. Banks should map these journeys and offer extensive customer touchpoints and compelling reasons for customers to move to the next stage. AI-powered voice bot or chatbot helps familiarize the customers with banking products and services and remove friction in banking journeys, thus driving new revenues and deepening relationships. Offer personalized products and services Banking enterprises are sitting on a treasure trove of customer data. Utilizing this data to offer more personalized customer service removes friction and delivers a transformed customer experience. Credit card companies have for long benefited by providing targeted services to their customers. Leveraging AI and predictive analytics, demographic-specific offers, location-based discounts, and brand loyalty benefits are other ways credit card providers reap rich benefits. Listen to the voice of the customer Gone are the days of one-size-fits-all banking services. Customers today have their primary goals laid bare, and then they have specific goals nested within these larger goals. Banks need to pay special attention to what exactly the customers are seeking. AI-based platforms like Conversational Service Automation help deliver a superior banking experience, thus empowering banks to deepen relationships with their customers. What is the customer’s implied financial needs, service expectations, apprehensions, unspoken priorities, ‘wow moment’ triggers, etc.? Machines can pore over the customer conversations across all channels for nuggets of information using conversational analytics. This way, you not only reduce the cost of customer acquisition but pave the way for a long-term relationship with banking customers based on mutual understanding.

In an attempt to earn customer loyalty heading into the Super Bowl pizza rush, Domino's is offering members of its Points for Pies rewards program a chance to earn points without having to buy a pizza from the chain. Domino's Senior Vice President and Chief Brand Officer Art D'Elia said in an announcement for the offering, "Instead of advertising during Sunday's game, we decided to invest in a breakthrough program." To receive the promotion, diners download the QSR's app and sign up for its rewards program. They can then use a "newly embedded pizza identification feature to scan their pizza" and earn 10 points from the company. Customers then redeem 60 points for a medium two-topping pizza at the restaurant chain.

Nikki Baird is Vice President of Retail Innovation at Aptos, a retail enterprise solution provider, as well co-founder of Retail Systems Research and a former analyst at Forrester Research. She discusses advancements in Artificial Intelligence that will help retailers ensure that AI isn't making bad assumptions under the adage "garbage in, garbage out" as well as the trend toward in-store personalization in Part 7 of EcommerceBytes Online Selling Trends 2019. Retail Breaks Out of the AI Black Box First-generation AI solutions were simple – data in, answer out. Solutions were designed to protect the average end user from confusion and distraction. While black box solutions serve their purpose, they also limit the value organizations can extrapolate by hiding AI logic, which in theory could be used to teach humans what was learned that led to various recommendations.

Bots are a great way for holiday shoppers to check off items from their wish lists while avoiding crowded stores and endless hold times on the phone. Bots are also a fantastic tool for the organizations that use them. But they might not be ready to help you find that perfect gift (yet). Internet bots execute routine online tasks, usually those either so tedious that they would bore human workers half to death, or on such a massive scale that people couldn't finish the job within their lifetime. But like people, bots learn -- and some even have personalities.

Subscribe to Datacenter and access the Agency Report 2016, tracking revenue and growth rates for more than 1K agencies, agency networks and agency companies.

The increasing number of genome-wide assays of gene expression available from public databases presents opportunities for computational methods that facilitate hypothesis generation and biological interpretation of these data. We present an unsupervised machine learning approach, ADAGE (analysis using denoising autoencoders of gene expression), and apply it to the publicly available gene expression data compendium for Pseudomonas aeruginosa. In this approach, the machine-learned ADAGE model contained 50 nodes which we predicted would correspond to gene expression patterns across the gene expression compendium. While no biological knowledge was used during model construction, cooperonic genes had similar weights across nodes, and genes with similar weights across nodes were significantly more likely to share KEGG pathways. By analyzing newly generated and previously published microarray and transcriptome sequencing data, the ADAGE model identified differences between strains, modeled the cellular response to low oxygen, and predicted the involvement of biological processes based on low-level gene expression differences. ADAGE compared favorably with traditional principal component analysis and independent component analysis approaches in its ability to extract validated patterns, and based on our analyses, we propose that these approaches differ in the types of patterns they preferentially identify. We provide the ADAGE model with analysis of all publicly available P. aeruginosa GeneChip experiments and open source code for use with other species and settings. Extraction of consistent patterns across large-scale collections of genomic data using methods like ADAGE provides the opportunity to identify general principles and biologically important patterns in microbial biology. This approach will be particularly useful in less-well-studied microbial species.