The debate around misinformation and its potentially detrimental effects on public opinion is complex and multifaceted, to the extent that even the relevant academic research has not found unanimity on the prevalence and consumption of misinformation compared with mainstream content. The methodological framework presented here emphasises the importance of considering data representative of the complexity of the phenomenon and metrics that control for possible scale effects. By combining statistical, econometric and machine learning models, we shed light on the real impact of misinformation about a subject of general interest and social relevance, such as vaccines, on both the information available to citizens and their news diet. Our results show the prominent role achieved by misinformation sources in the news ecosystem, but also - and above all - the inability of mainstream media to drive the public debate over time on issues that are particularly sensitive and emotional. Taking properly account for the temporal dynamics of public debate seems crucial to prevent the latter from moving into uncontrolled spaces where false narratives are more easily conveyed and entrenched.

Recommender systems are a long-standing research problem in data mining and machine learning. They are incremental in nature, as new user-item interaction logs arrive. In real-world applications, we need to periodically train a collaborative filtering algorithm to extract user/item embedding vectors and therefore, a time-series of embedding vectors can be naturally defined. We present a time-series forecasting-based upgrade kit (TimeKit), which works in the following way: it i) first decides a base collaborative filtering algorithm, ii) extracts user/item embedding vectors with the base algorithm from user-item interaction logs incrementally, e.g., every month, iii) trains our time-series forecasting model with the extracted time-series of embedding vectors, and then iv) forecasts the future embedding vectors and recommend with their dot-product scores owing to a recent breakthrough in processing complicated time-series data, i.e., neural controlled differential equations (NCDEs). Our experiments with four real-world benchmark datasets show that the proposed time-series forecasting-based upgrade kit can significantly enhance existing popular collaborative filtering algorithms.

We report on novel investigations into training models that make sentences concise. We define the task and show that it is different from related tasks such as summarization and simplification. For evaluation, we release two test sets, consisting of 2000 sentences each, that were annotated by two and five human annotators, respectively. We demonstrate that conciseness is a difficult task for which zero-shot setups with large neural language models often do not perform well. Given the limitations of these approaches, we propose a synthetic data generation method based on round-trip translations. Using this data to either train Transformers from scratch or fine-tune T5 models yields our strongest baselines that can be further improved by fine-tuning on an artificial conciseness dataset that we derived from multi-annotator machine translation test sets.

In the early '80s, that elusive dream manifested as Laserdisc interactive cartoons (such as Konami's spectacular 1984 Western, "Badlands") and VHS-powered, live-action shooting galleries (developed for experimental home consoles, like the Action Max). Both approaches ensured a measure of spectacle light-years ahead of anything offered by the home consoles of the era, but the price for their visual extravagance was rigid, predetermined action sequences and a lack of meaningful engagement. Players would simply point plastic guns at their TV set to rack up points or wait for the prompt to perform button presses that had little to do with what happened on the screen -- early precursors to contemporary quick time events.

Netflix officially launched its £4.99-a-month subscription service in the UK last week, with 30 second adverts shown during films and TV shows. 'Basic with Adverts' is £11 cheaper each month than the streamer's most expensive package, as viewers will be subjected to four to five minutes of adverts per hour. And there's another catch - many of Netflix's most popular movies and programmes are blocked from this option. This includes the sitcoms New Girl and Brooklyn Nine-Nine, dramas Suits and House of Cards, and movies Paddington and The Imitation Game. Netflix explained on its website: 'Some TV shows and movies aren't available to watch with the Basic with ads plan because of licensing restrictions. 'These titles will have a lock icon when you search or browse Netflix.'

AI art has already proven itself to be pretty scary, for a number a reasons. Not only are image generators like DALL-E 2 capable of churning out some pretty nightmarish images, but there's also the ethical and existential question around what the rise of automated art might mean for artists. Leaving the latter question aside for now, here's some art that takes the idea of terrifying AI art to its logical conclusion. Digital agency Evoluted has submitted horror film-related keywords into an AI tool to create alternative posters for various classics – and the results are scarier than the originals. What to give it a go for yourself?

NLP Logix, an artificial intelligence and machine learning solutions company, today announced it's partnership and inclusion into the Beeline ecosystem. Beeline customers will be able to leverage NLP Logix's task automation solutions to drive even greater efficiencies. Beeline is the leading technology solution provider for managing the global extended workforce. "NLP Logix's AI tools can eliminate repetitive tasks, allowing the contingent workforce to offer even more value." "Beeline's mission is to optimize the contingent workforce in every way and partnering with NLP Logix will add another layer to this optimization," said Doug Leeby, Beeline CEO.

CMU and MIT Researchers Developed a Machine Learning System Called Neural Acoustic Fields (NAFs) That can Simulate How a Listener Would Hear a …

Decision making systems trained with real-world data are deployed ubiquitously in our daily life, for example, in regard to credit, education, and medical care. However, those decision systems may demonstrate discrimination against disadvantaged groups due to the biases in the data [16]. In order to mitigate this issue, many have proposed to impose fairness constraints [16, 20] on the decision, such that certain statistical parity properties are achieved. Despite the fact that fair learning has been extensively studied, most of this work is in the static setting without considering the sequential feedback effects of decisions. At the same time, in many scenarios, algorithmic decisions may incur changes in the underlying features or qualification status of individuals, which further feeds back to the decision making process; for example, banks' decision may induce borrowers to react, for example changing their FICO score by closing credit cards.

The heterogeneity gap problem is the main challenge in cross-modal retrieval. Because cross-modal data (e.g. audiovisual) have different distributions and representations that cannot be directly compared. To bridge the gap between audiovisual modalities, we learn a common subspace for them by utilizing the intrinsic correlation in the natural synchronization of audio-visual data with the aid of annotated labels. TNN-CCCA is the best audio-visual cross-modal retrieval (AV-CMR) model so far, but the model training is sensitive to hard negative samples when learning common subspace by applying triplet loss to predict the relative distance between inputs. In this paper, to reduce the interference of hard negative samples in representation learning, we propose a new AV-CMR model to optimize semantic features by directly predicting labels and then measuring the intrinsic correlation between audio-visual data using complete cross-triple loss. In particular, our model projects audio-visual features into label space by minimizing the distance between predicted label features after feature projection and ground label representations. Moreover, we adopt complete cross-triplet loss to optimize the predicted label features by leveraging the relationship between all possible similarity and dissimilarity semantic information across modalities. The extensive experimental results on two audio-visual double-checked datasets have shown an improvement of approximately 2.1% in terms of average MAP over the current state-of-the-art method TNN-CCCA for the AV-CMR task, which indicates the effectiveness of our proposed model.