Visual question answering (VQA) is a critical multimodal task in which an agent must answer questions according to the visual cue. Unfortunately, language bias is a common problem in VQA, which refers to the model generating answers only by associating with the questions while ignoring the visual content, resulting in biased results. We tackle the language bias problem by proposing a self-supervised counterfactual metric learning (SC-ML) method to focus the image features better. SC-ML can adaptively select the question-relevant visual features to answer the question, reducing the negative influence of question-irrelevant visual features on inferring answers. In addition, question-irrelevant visual features can be seamlessly incorporated into counterfactual training schemes to further boost robustness. Extensive experiments have proved the effectiveness of our method with improved results on the VQA-CP dataset. Our code will be made publicly available.

Brazil, India, China, and South Africa are the only countries where more than half of the population expressed strong trust and acceptance of artificial intelligence technologies, according to a study from global accounting firm KPMG. The country with the highest trust in A.I. is India, with a 75% overall acceptance rate. Moreover, the study revealed that emerging countries --specifically the BRICS bloc-- also have the highest engagement with A.I. China is the nation with the most people using A.I. in their workplace (75%), followed by India with 66% and Brazil with 50%. On the other hand, citizens of developed countries appeared to be more skeptical.

Wild elephants are one of the few known species to show signs of self-domestication. The phenomenon has only previously been documented in humans and bonobos, a closely related primate. Humans have bred animals to maximise traits such as friendliness, sociability and a docile temperament in a process called domestication. Some researchers believe humans and bonobos have gone through a similar process but that they have naturally done it to themselves. Limor Raviv at the Max Planck Institute for Psycholinguistics in the Netherlands wondered if other species have self-domesticated too.

LONDON, April 03, 2023--(BUSINESS WIRE)--C3 AI (NYSE: AI), the Enterprise AI application software company, has today announced the relocation of their European, Middle East and Africa (EMEA) headquarters from Paris, France to London, United Kingdom. "Our new EMEA headquarters represents an important next step for C3 AI," said C3 AI CEO Thomas M. Siebel. "As a global company, we are proud to call London our EMEA base and we look forward to the opportunity for even closer collaboration with the UK's world-class technology and artificial intelligence experts, businesses, and institutions." C3 AI has had a presence in Europe since the company's inception in 2009, with offices currently located in London, Paris, Munich, Rome, and Amsterdam. The decision to relocate the central EMEA operation to London ties in with the business's ambitious 2023 targets.

The star of Levi's new campaign looks like any other model . Her tousled hair hangs over her shouldersas she gazes into the camera with that far-off high-fashion stare. But look closer, and something starts to seem a little off. The shadow between her chin and neck looks muddled, like a bad attempt at using FaceTune's eraser effect to hide a double chin. Her French manicured fingernails appear scrubbed clean and uniform in a creepy real doll kind of way.

Hosted by Eleanor Drage and Kerry Mackereth, The Good Robot is a podcast which explores the many complex intersections between gender, feminism and technology. Bridget is an expert in how AI is impacting Africa and the major challenges in implementing AI use across the continent. She tells us about what good technology means in the contexts in which she works and the benefits and drawbacks of Google and other Big Tech companies operating in Africa. Bridget Boakye is a Ghanaian entrepreneur, data scientist and writer. She is.the Artificial Intelligence Lead in the Internet Policy Unit of the Tony Blair Institute.

Crayon Data, an artificial intelligence and big data company headquartered in Singapore with a presence in Asia, Middle East and Africa, has announced its partnership with Visa. The partnership will empower Visa's global issuers to create digital experiences for their customers with AI-led relevance and personalisation solutions. Crayon Data's flagship platform, maya.ai, "Visa has always embraced innovation at its core. With cashless payments taking the driver's seat, APAC is set to grow fastest in cashless transaction volume growing by 109%, followed by Africa and Europe. Visa's vision to foster financial inclusion across the globe, will make this partnership the benchmark in the payments ecosystem to accelerate the creation of compelling digital experiences for consumers."

We demonstrate when and how an entire left-infinite orbit of an underlying dynamical system or observations from such left-infinite orbits can be uniquely represented by a pair of elements in a different space, a phenomenon which we call \textit{causal embedding}. The collection of such pairs is derived from a driven dynamical system and is used to learn a function which together with the driven system would: (i). determine a system that is topologically conjugate to the underlying system (ii). enable forecasting the underlying system's dynamics since the conjugacy is computable and universal, i.e., it does not depend on the underlying system (iii). guarantee an attractor containing the image of the causally embedded object even if there is an error made in learning the function. By accomplishing these we herald a new forecasting scheme that beats the existing reservoir computing schemes that often lead to poor long-term consistency as there is no guarantee of the existence of a learnable function, and overcomes the challenges of stability in Takens delay embedding. We illustrate accurate modeling of underlying systems where previously known techniques have failed.

Referring image segmentation (RIS) aims to find a segmentation mask given a referring expression grounded to a region of the input image. Collecting labelled datasets for this task, however, is notoriously costly and labor-intensive. To overcome this issue, we propose a simple yet effective zero-shot referring image segmentation method by leveraging the pre-trained cross-modal knowledge from CLIP. In order to obtain segmentation masks grounded to the input text, we propose a mask-guided visual encoder that captures global and local contextual information of an input image. By utilizing instance masks obtained from off-the-shelf mask proposal techniques, our method is able to segment fine-detailed Istance-level groundings. We also introduce a global-local text encoder where the global feature captures complex sentence-level semantics of the entire input expression while the local feature focuses on the target noun phrase extracted by a dependency parser. In our experiments, the proposed method outperforms several zero-shot baselines of the task and even the weakly supervised referring expression segmentation method with substantial margins. Our code is available at https://github.com/Seonghoon-Yu/Zero-shot-RIS.

Studies on benign overfitting provide insights for the success of overparameterized deep learning models. In this work, we examine whether overfitting is truly benign in real-world classification tasks. We start with the observation that a ResNet model overfits benignly on Cifar10 but not benignly on ImageNet. To understand why benign overfitting fails in the ImageNet experiment, we theoretically analyze benign overfitting under a more restrictive setup where the number of parameters is not significantly larger than the number of data points. Under this mild overparameterization setup, our analysis identifies a phase change: unlike in the previous heavy overparameterization settings, benign overfitting can now fail in the presence of label noise. Our analysis explains our empirical observations, and is validated by a set of control experiments with ResNets. Our work highlights the importance of understanding implicit bias in underfitting regimes as a future direction.