We knew Amazon would revive its Prime Big Deal Days sale event this fall, but we didn't know the exact dates until today. The online retailer announced that the sale event will return this year on October 8 and 9, giving us all the more reason to call it October Prime Day as we have done in years past. This is the third iteration of the fall sale event that Amazon has used as its (un)official kickoff to the holiday shopping season. Prime Day in July remains Amazon's marquee sale event for Prime members, but ever since its debut in 2022, October Prime Day provides subscribers with thousands of exclusive deals to shop during the two-day window. In turn, it also provides Amazon a way to boost sales during the same time period and, arguably more importantly, increase the number of overall Prime subscribers.

Debiasing is often challenging due to computational costs, data constraints, and potential degradation of multi-task language capabilities. This work introduces a novel approach utilizing ChatGPT to generate synthetic training data, aiming to enhance the debiasing of LLMs. We propose two strategies: Targeted Prompting, which provides effective debiasing for known biases but necessitates prior specification of bias in question; and General Prompting, which, while slightly less effective, offers debiasing across various categories. We leverage resource-efficient LLM debiasing using adapter tuning and compare the effectiveness of our synthetic data to existing debiasing datasets. Our results reveal that: (1) ChatGPT can efficiently produce high-quality training data for debiasing other LLMs; (2) data produced via our approach surpasses existing datasets in debiasing performance while also preserving internal knowledge of a pre-trained LLM; and (3) synthetic data exhibits generalizability across categories, effectively mitigating various biases, including intersectional ones. These findings underscore the potential of synthetic data in advancing the fairness of LLMs with minimal retraining cost.

Motivated by online retail, we consider the problem of selling one item (e.g., an ad slot) to two non-excludable buyers (say, a merchant and a brand). This problem captures, for example, situations where a merchant and a brand cooperatively bid in an auction to advertise a product, and both benefit from the ad being shown. A mechanism collects bids from the two and decides whether to allocate and which payments the two parties should make. This gives rise to intricate incentive compatibility constraints, e.g., on how to split payments between the two parties. We approach the problem of finding a revenue-maximizing incentive-compatible mechanism from an online learning perspective; this poses significant technical challenges. First, the action space (the class of all possible mechanisms) is huge; second, the function that maps mechanisms to revenue is highly irregular, ruling out standard discretization-based approaches. In the stochastic setting, we design an efficient learning algorithm achieving a regret bound of $O(T^{3/4})$. Our approach is based on an adaptive discretization scheme of the space of mechanisms, as any non-adaptive discretization fails to achieve sublinear regret. In the adversarial setting, we exploit the non-Lipschitzness of the problem to prove a strong negative result, namely that no learning algorithm can achieve more than half of the revenue of the best fixed mechanism in hindsight. We then consider the $\sigma$-smooth adversary; we construct an efficient learning algorithm that achieves a regret bound of $O(T^{2/3})$ and builds on a succinct encoding of exponentially many experts. Finally, we prove that no learning algorithm can achieve less than $\Omega(\sqrt T)$ regret in both the stochastic and the smooth setting, thus narrowing the range where the minimax regret rates for these two problems lie.

Binary human choice feedback is widely used in interactive preference learning for its simplicity, but it provides limited information about preference strength. To overcome this limitation, we leverage human response times, which inversely correlate with preference strength, as complementary information. Our work integrates the EZ-diffusion model, which jointly models human choices and response times, into preference-based linear bandits. We introduce a computationally efficient utility estimator that reformulates the utility estimation problem using both choices and response times as a linear regression problem. Theoretical and empirical comparisons with traditional choice-only estimators reveal that for queries with strong preferences ("easy" queries), choices alone provide limited information, while response times offer valuable complementary information about preference strength. As a result, incorporating response times makes easy queries more useful. We demonstrate this advantage in the fixed-budget best-arm identification problem, with simulations based on three real-world datasets, consistently showing accelerated learning when response times are incorporated.

Large Language Models (LLMs) have shown versatility in various Natural Language Processing (NLP) tasks, including their potential as effective question-answering systems. However, to provide precise and relevant information in response to specific customer queries in industry settings, LLMs require access to a comprehensive knowledge base to avoid hallucinations. Retrieval Augmented Generation (RAG) emerges as a promising technique to address this challenge. Yet, developing an accurate question-answering framework for real-world applications using RAG entails several challenges: 1) data availability issues, 2) evaluating the quality of generated content, and 3) the costly nature of human evaluation. In this paper, we introduce an end-to-end framework that employs LLMs with RAG capabilities for industry use cases. Given a customer query, the proposed system retrieves relevant knowledge documents and leverages them, along with previous chat history, to generate response suggestions for customer service agents in the contact centers of a major retail company. Through comprehensive automated and human evaluations, we show that this solution outperforms the current BERT-based algorithms in accuracy and relevance. Our findings suggest that RAG-based LLMs can be an excellent support to human customer service representatives by lightening their workload.

John Lewis brings back'never knowingly undersold' Getty Images Retailer John Lewis is bringing back its "never knowingly undersold" price pledge from Monday, two years after abandoning it. It will also apply to online sales for the first time, whereas it previously only applied to in-store shopping, and will use AI to match the prices of 25 top retailers. The department store chain has been trying to win back customers after a tough few years that has seen it cut jobs and close several stores. It swung back to profit earlier this year, but is expected to continue shedding jobs as it seeks to revive its fortunes. The decision by John Lewis' new managing director Pete Ruis to restore the price pledge marks a change of direction from his predecessor.

Marks & Spencer is using artificial intelligence to advise shoppers on their outfit choices based on their body shape and style preferences, as part of efforts to increase online sales. The 130-year-old retailer is using the technology to personalise consumers' online experience, and suggest items to buy. Stephen Langford, the company's director of online, said M&S was using AI to adapt the language used to address shoppers, tailoring to six different preferences such as emotional, descriptive language or more straightforward prose. One of its aims is to personalise online interactions with shoppers, he said, such as prioritising products most relevant for an individual. Male shoppers would be less likely to be offered the latest deals on bras, for example.

The humble jar of honey might seem sweet and innocent, but experts warn that British shoppers have been getting stung when spending on this staple. Investigations have revealed that unscrupulous honey producers around the world bulk out their products with cheap sugars that are almost impossible to detect. However, scientists have now developed a test which can easily spot the difference between fake and real honey - without even opening the jar. The light-based technique can detect the unique chemical signature of real honey as well as the syrups that try to imitate it. While the test isn't readily available yet, experts told MailOnline that consumers may be able to spot the frauds in their cupboards using nothing more than their phone torch within five to 10 years.

Save big on back to school and dorm room essentials on Amazon. During the back-to-school shopping season, you can find deep discounts on top brands on Amazon. Now is your chance to stock up on school supplies, backpacks and dorm room essentials for up to 50% off the list price. Get your back-to-school discounts delivered in time for the first day by signing up for a Prime membership. The benefits include fast, free delivery, access to invite-only deals and the option to Buy With Prime. Most purchases can be delivered to your door in 24 hours if you're an Amazon Prime member.

$K$-nearest neighbor language models ($k$NN-LMs), which integrate retrieval with next-word prediction, have demonstrated strong performance in language modeling as well as downstream NLP benchmarks. These results have led researchers to argue that models trained on poor quality or outdated data could perform well by employing a $k$NN extension that has access to a higher-quality datastore. In this work, we ask whether this improved ability to recall information really translates into downstream abilities. We extensively evaluate $k$NN-LMs on a diverse set of tasks, ranging from sentiment classification and commonsense reasoning to multi-hop reasoning. Results show that $k$NN-LMs excel at memory-intensive tasks, where utilizing the patterns in the input is sufficient for determining the output, but struggle with reasoning tasks that require integrating multiple pieces of information to derive new knowledge. We further demonstrate through oracle experiments and qualitative analysis that even with perfect retrieval, $k$NN-LMs still fail to determine the correct answers, placing an upper bound on their reasoning performance. Code and datastores are released at https://github.com/GSYfate/knnlm-limits/.