Unscrupulous foreign firms are using AI-generated images and false back stories to pose as family-run UK businesses to lure in shoppers. Customers say they feel completely ripped off after believing they were buying from independent boutiques in England but were delivered cheap clothes and jewellery, mass-shipped from warehouses in east Asia. Among the websites is C'est La Vie, a shop purporting to be run by couple Eileen and Patrick for 29 years and based in Birmingham's historic Jewellery Quarter - but with a returns address in China. Consumer guide Which? said the growing use of AI tools was making it possible for fraudsters to mislead the public on an unprecedented scale. Another website appearing to use AI-generated images is Mabel & Daisy, a seemingly quintessential, mother and daughter-owned clothing firm, which claims to be based in Bristol but has an address in Hong Kong.

Natural language processing (NLP) models may leak private information in different ways, including membership inference, reconstruction or attribute inference attacks. Sensitive information may not be explicit in the text, but hidden in underlying writing characteristics. Methods to protect privacy can involve using representations inside models that are demonstrated not to detect sensitive attributes or -- for instance, in cases where users might not trust a model, the sort of scenario of interest here -- changing the raw text before models can have access to it. The goal is to rewrite text to prevent someone from inferring a sensitive attribute (e.g. the gender of the author, or their location by the writing style) whilst keeping the text useful for its original intention (e.g. the sentiment of a product review). The few works tackling this have focused on generative techniques. However, these often create extensively different texts from the original ones or face problems such as mode collapse. This paper explores a novel adaptation of adversarial attack techniques to manipulate a text to deceive a classifier w.r.t one task (privacy) whilst keeping the predictions of another classifier trained for another task (utility) unchanged. We propose IDT, a method that analyses predictions made by auxiliary and interpretable models to identify which tokens are important to change for the privacy task, and which ones should be kept for the utility task. We evaluate different datasets for NLP suitable for different tasks. Automatic and human evaluations show that IDT retains the utility of text, while also outperforming existing methods when deceiving a classifier w.r.t privacy task.

Pre-trained language models (PLMs) have outperformed other NLP models on a wide range of tasks. Opting for a more thorough understanding of their capabilities and inner workings, researchers have established the extend to which they capture lower-level knowledge like grammaticality, and mid-level semantic knowledge like factual understanding. However, there is still little understanding of their knowledge of higher-level aspects of language. In particular, despite the importance of sociodemographic aspects in shaping our language, the questions of whether, where, and how PLMs encode these aspects, e.g., gender or age, is still unexplored. We address this research gap by probing the sociodemographic knowledge of different single-GPU PLMs on multiple English data sets via traditional classifier probing and information-theoretic minimum description length probing. Our results show that PLMs do encode these sociodemographics, and that this knowledge is sometimes spread across the layers of some of the tested PLMs. We further conduct a multilingual analysis and investigate the effect of supplementary training to further explore to what extent, where, and with what amount of pre-training data the knowledge is encoded. Our overall results indicate that sociodemographic knowledge is still a major challenge for NLP. PLMs require large amounts of pre-training data to acquire the knowledge and models that excel in general language understanding do not seem to own more knowledge about these aspects.

The accelerated growth in ecommerce and online marketplaces has led to a surge in fraudulent behavior online perpetrated by bots and bad actors alike. A strategic and effective approach to online fraud detection will be needed in order to tackle increasingly sophisticated threats to online retailers. These market shifts come at a time of significant regulatory change. Across the globe, new legislation is coming into force that alters the balance of responsibility in fraud prevention between users, brands, and the platforms that promote them digitally. For example, the EU Digital Services Act and US Shop Safe Act will require online platforms to take greater responsibility for the content on their websites, a responsibility that was traditionally the domain of brands and users to monitor and report.

This started out as a challenge. I wanted with a friend of mine to see if it was possible to build something from scratch and push it to production. In this post, we'll go through the necessary steps to build and deploy a machine learning application. This starts from data collection to deployment and the journey, as you'll see it, is exciting and fun . Before we begin, let's have a look at the app we'll be building: As you see, this web app allows a user to evaluate random brands by writing reviews.

This started out as a challenge. I wanted with a friend of mine to see if it was possible to build something from scratch and push it to production. In this post, we'll go through the necessary steps to build and deploy a machine learning application. This starts from data collection to deployment and the journey, as you'll see it, is exciting and fun . Before we begin, let's have a look at the app we'll be building: As you see, this web app allows a user to evaluate random brands by writing reviews.

Salesforce has officially unveiled Einstein, a set of artificial intelligence (AI) capabilities it says will help users of its platform serve their customers better. Billing the technology as "AI for everyone", Salesforce is putting Einstein's capabilities into all its clouds, bringing machine learning, deep learning, predictive analytics, and natural language processing into each piece of its customer relationship management platform. In Salesforce's Sales Cloud, for instance, machine learning will power predictive lead scoring, a new tool that can analyse all data related to leads -- including standard and custom fields, activity data from sales reps, and behavioural activity from prospects -- to generate a predictive score for each lead. The models will continuously improve over time by learning from signals like lead source, industry, job title, Web clicks and emails. Another tool will analyse CRM data combined with customer interactions such as inbound emails from prospects to identify buying signals earlier in the sales process and recommend next steps to increase the sales rep's ability to close a deal.