Conversational AI systems can engage in unsafe behaviour when handling users' medical queries that can have severe consequences and could even lead to deaths. Systems therefore need to be capable of both recognising the seriousness of medical inputs and producing responses with appropriate levels of risk. We create a corpus of human written English language medical queries and the responses of different types of systems. We label these with both crowdsourced and expert annotations. While individual crowdworkers may be unreliable at grading the seriousness of the prompts, their aggregated labels tend to agree with professional opinion to a greater extent on identifying the medical queries and recognising the risk types posed by the responses. Results of classification experiments suggest that, while these tasks can be automated, caution should be exercised, as errors can potentially be very serious.

Artificial intelligence (AI) has evolved into a personal digital assistant rather than a physical machine like the Terminator, says Samer Shoueiry of the creative communications group Publicis Communications. And it is developing at a speed that is "incredible", adds the regional executive head of digital and social marketing for the company. "We see AI today in everything we do as the personal digital assistant – someone returning your web searches, filtering your spam emails, the technology assisting everything within your car," he says. AI can be seen in every aspect of technology that links to digitisation, adds Mr Shoueiry, who spoke at the recent Dubai Lynx advertising festival. Research company Forrester claims that cognitive technologies such as AI and machine learning will replace 16 per cent of US jobs by 2025, in a report it published last year entitled Sharing Your Cubicle With Robots.

Days after hosting a major hardware launch, Amazon is apparently having a sale on its older Echo devices. Among the deals, we noticed that both the Echo Show 5 and the larger Echo Show 8 have hit new record lows. The Echo Show 5, which went on sale last year for $85, is now down to $35, a 59 percent discount. The 8-inch model, meanwhile, is down to $70 after having debuted at $100. Both devices, but especially the Echo Show 8, were designed to be used as a possible alarm clock, with a sunrise alarm feature that gently wakes you up by slowly brightening the display.

Executives across industries are turning to automation to deliver on cost optimization and enhanced productivity objectives, Saikat Ray, VP analyst at Gartner, told CIO Dive in August. The robotic process automation software market will reach $2.9 billion by the end of 2022, up 19.5% from 2021, according to Gartner. In recent years, large UiPath bank customers have been using automation tools to facilitate initiatives that include data extraction and data transfer efforts to support the merger of BB&T and SunTrust; reduction of manual work for Wells Fargo contact center agents through digital personal assistants; and the delegation of some structured, rule-based repetitive tasks to bots at JPMorgan Chase. From JPMorgan Chase's perspective, one of the next steps on its automation journey will include using bots to tackle more sophisticated tasks, including delving into unstructured processes and unstructured data, and using machine learning to facilitate these efforts, said Shefali Shah, managing director of global digital transformation and integrated intelligent automation at JPMorgan Chase. Diana Caplinger, executive vice president and head of enterprise enablement and intelligent automation at Truist, said the company is deploying automation in support of "integrated relationship management," an effort to use data across the organization to deliver more personalized service to clients.

Conversational AI has advanced dramatically since the early days of chatbots with limited capabilities. As time goes on, more industries are realizing the capabilities and benefits these advancements bring. Digital assistants like Alexa and Siri have consumers wondering why the same capabilities can't be used at work. While there are enterprise versions of Alexa and Cortana, conversational AI is still not at a point where a user can ask any question and receive a coherent answer. Like most other types of AI, the best use cases are narrow as opposed to broad.

We showcase a novel solution to a recommendation system problem where we face a perpetual soft item cold start issue. Our system aims to recommend demanded products to prospective sellers for listing in Amazon stores. These products always have only few interactions thereby giving rise to a perpetual soft item cold start situation. Modern collaborative filtering methods solve cold start using content attributes and exploit the existing implicit signals from warm start items. This approach fails in our use-case since our entire item set faces cold start issue always. Our Product Graph has over 500 Million nodes and over 5 Billion edges which makes training and inference using modern graph algorithms very compute intensive. To overcome these challenges we propose a system which reduces the dataset size and employs an improved modelling technique to reduce storage and compute without loss in performance. Particularly, we reduce our graph size using a filtering technique and then exploit this reduced product graph using Weighted Averaging of Messages over Layers (WAML) algorithm. WAML simplifies training on large graphs and improves over previous methods by reducing compute time to 1/7 of LightGCN and 1/26 of Graph Attention Network (GAT) and increasing recall$@100$ by 66% over LightGCN and 2.3x over GAT.

Online stores and service providers rely heavily on recommendation softwares to guide users through the vast amount of available products. Consequently, the field of recommender systems has attracted increased attention from the industry and academia alike, but despite this joint effort, the field still faces several challenges. For instance, most existing work models the recommendation problem as a matrix completion problem to predict the user preference for an item. This abstraction prevents the system from utilizing the rich information from the ordered sequence of user actions logged in online sessions. To address this limitation, researchers have recently developed a promising new breed of algorithms called sequence-aware recommender systems to predict the user's next action by utilizing the time series composed of the sequence of actions in an ongoing user session. This paper proposes a novel sequence-aware recommendation approach based on a complex network generated by the hidden metric space model, which combines node similarity and popularity to generate links. We build a network model from data and then use it to predict the user's subsequent actions. The network model provides an additional source of information that improves the accuracy of the recommendations. The proposed method is implemented and tested experimentally on a large dataset. The results prove that the proposed approach performs better than state-of-the-art recommendation methods.

Voice user interfaces and digital assistants are rapidly entering our lives and becoming singular touch points spanning our devices. These always-on services capture and transmit our audio data to powerful cloud services for further processing and subsequent actions. Our voices and raw audio signals collected through these devices contain a host of sensitive paralinguistic information that is transmitted to service providers regardless of deliberate or false triggers. As our emotional patterns and sensitive attributes like our identity, gender, well-being, are easily inferred using deep acoustic models, we encounter a new generation of privacy risks by using these services. One approach to mitigate the risk of paralinguistic-based privacy breaches is to exploit a combination of cloud-based processing with privacy-preserving, on-device paralinguistic information learning and filtering before transmitting voice data. In this paper we introduce EDGY, a configurable, lightweight, disentangled representation learning framework that transforms and filters high-dimensional voice data to identify and contain sensitive attributes at the edge prior to offloading to the cloud. We evaluate EDGY's on-device performance and explore optimization techniques, including model quantization and knowledge distillation, to enable private, accurate and efficient representation learning on resource-constrained devices. Our results show that EDGY runs in tens of milliseconds with 0.2% relative improvement in "zero-shot" ABX score or minimal performance penalties of approximately 5.95% word error rate (WER) in learning linguistic representations from raw voice signals, using a CPU and a single-core ARM processor without specialized hardware.

Collaborative Metric Learning (CML) has recently emerged as a popular method in recommendation systems (RS), closing the gap between metric learning and Collaborative Filtering. Following the convention of RS, existing methods exploit unique user representation in their model design. This paper focuses on a challenging scenario where a user has multiple categories of interests. Under this setting, we argue that the unique user representation might induce preference bias, especially when the item category distribution is imbalanced. To address this issue, we propose a novel method called \textit{Diversity-Promoting Collaborative Metric Learning} (DPCML), with the hope of considering the commonly ignored minority interest of the user. The key idea behind DPCML is to include a multiple set of representations for each user in the system. Based on this embedding paradigm, user preference toward an item is aggregated from different embeddings by taking the minimum item-user distance among the user embedding set. Furthermore, we observe that the diversity of the embeddings for the same user also plays an essential role in the model. To this end, we propose a \textit{diversity control regularization} term to accommodate the multi-vector representation strategy better. Theoretically, we show that DPCML could generalize well to unseen test data by tackling the challenge of the annoying operation that comes from the minimum value. Experiments over a range of benchmark datasets speak to the efficacy of DPCML.

Alexa, Siri and Google Home might be making children less intelligent and socially stunted, it was claimed today. The voice-controlled devices -- popular in homes across the world -- allow users to ask questions and receive answers. But this may impede youngster's learning skills, critical thinking and empathy, says Dr Anmol Arora, a researcher at Cambridge University. Dr Anmol Arora, a researcher at Cambridge University, says this is down to the tech only offering short and concise answers to questions, inappropriate responses and being unable to give feedback on their social skills. Alexa, Siri and Google Home might be making children less intelligent and socially stunted, according to an artificial intelligence expert.