Beth Barnes and three of her colleagues sit cross-legged in a semicircle on a damp lawn on the campus of the University of California, Berkeley. They are describing their attempts to interrogate artificial intelligence chatbots. "They are, in some sense, these vast alien intelligences," says Barnes, 26, who is the founder and CEO of Model Evaluation and Threat Research (METR), an AI-safety nonprofit. "They know so much about whether the next word is going to be'is' versus'was.' We're just playing with a tiny bit on the surface, and there's all this, miles and miles underneath," she says, gesturing at the potentially immense depths of large language models' capabilities. Researchers at METR look a lot like Berkeley students--the four on the lawn are in their twenties and dressed in jeans or sweatpants.

In deductive domains, three metacognitive knowledge types in ascending order are declarative, procedural, and conditional learning. This work leverages Deep Reinforcement Learning (DRL) in providing adaptive metacognitive interventions to bridge the gap between the three knowledge types and prepare students for future learning across Intelligent Tutoring Systems (ITSs). Students received these interventions that taught how and when to use a backward-chaining (BC) strategy on a logic tutor that supports a default forward-chaining strategy. Six weeks later, we trained students on a probability tutor that only supports BC without interventions. Our results show that on both ITSs, DRL bridged the metacognitive knowledge gap between students and significantly improved their learning performance over their control peers. Furthermore, the DRL policy adapted to the metacognitive development on the logic tutor across declarative, procedural, and conditional students, causing their strategic decisions to be more autonomous.

This work compares two approaches to provide metacognitive interventions and their impact on preparing students for future learning across Intelligent Tutoring Systems (ITSs). In two consecutive semesters, we conducted two classroom experiments: Exp. 1 used a classic artificial intelligence approach to classify students into different metacognitive groups and provide static interventions based on their classified groups. In Exp. 2, we leveraged Deep Reinforcement Learning (DRL) to provide adaptive interventions that consider the dynamic changes in the student's metacognitive levels. In both experiments, students received these interventions that taught how and when to use a backward-chaining (BC) strategy on a logic tutor that supports a default forward-chaining strategy. Six weeks later, we trained students on a probability tutor that only supports BC without interventions. Our results show that adaptive DRL-based interventions closed the metacognitive skills gap between students. In contrast, static classifier-based interventions only benefited a subset of students who knew how to use BC in advance. Additionally, our DRL agent prepared the experimental students for future learning by significantly surpassing their control peers on both ITSs.

Explainable artificial intelligence (XAI) methods shed light on the predictions of deep neural networks (DNNs). Several different approaches exist and have partly already been successfully applied in climate science. However, the often missing ground truth explanations complicate their evaluation and validation, subsequently compounding the choice of the XAI method. Therefore, in this work, we introduce XAI evaluation in the context of climate research and assess different desired explanation properties, namely, robustness, faithfulness, randomization, complexity, and localization. To this end we build upon previous work and train a multi-layer perceptron (MLP) and a convolutional neural network (CNN) to predict the decade based on annual-mean temperature maps. Next, multiple local XAI methods are applied and their performance is quantified for each evaluation property and compared against a baseline test. Independent of the network type, we find that the XAI methods Integrated Gradients, Layer-wise relevance propagation, and InputGradients exhibit considerable robustness, faithfulness, and complexity while sacrificing randomization. The opposite is true for Gradient, SmoothGrad, NoiseGrad, and FusionGrad. Notably, explanations using input perturbations, such as SmoothGrad and Integrated Gradients, do not improve robustness and faithfulness, contrary to previous claims. Overall, our experiments offer a comprehensive overview of different properties of explanation methods in the climate science context and supports users in the selection of a suitable XAI method.

Chang'e 5 is on the last leg of its mission on the moon. After a visit to the lunar surface lasting less than 48 hours, it is back in orbit around the moon and ready to bring its samples home so that scientists on Earth can analyse them. The spacecraft consists of an orbiter, re-entry capsule, a lander and ascent stage, and launched on 23 November aboard a Long March 5 rocket. It landed on the moon on 1 December. It is China's first sample return mission, making the nation only the third – after the US and the Soviet Union – to bring back rocks and dust from the moon.

NEW DELHI: Facebook on Tuesday said it has stepped up the use of artificial intelligence (AI) to " prioritise reported content", a move that will help the social media giant take action faster on harmful and violative content. Facebook, which has 1.82 billion daily users globally, has drawn flak in the past for its handling of hate speech on the platform in India, which is among its biggest markets. Facebook Product Manager (Community Integrity) Ryan Barnes said the company is using AI to prioritise reported content, and that this prioritisation is important to help its over 15,000 reviewers. She explained that the prioritisation is important for four reasons -- not all harmful content is equal, some enforcement decisions are complex, people do not always report harmful content and the reports aren't always accurate. Speaking to reporters in a virtual briefing, she said the company has moved from relying on user reports alone to add use of technology to help aid the process.

During a press meet recently, a Facebook spokesperson said that the social media giant would be redoubling its efforts to counter'harmful content' on its platform using artificial intelligence. Reportedly, Ryan Barnes, the Facebook Product Manager of Community Integrity, said that the company would use AI to prioritise harmful content. This move is targeting at helping its over 15,000 human reviewers and moderators in dealing with reported contents. Barnes said during the press interaction, "We want to make sure we're getting to the worst of the worst, prioritising real-world imminent harm above all." With that being said, there have been numerous attempts in the past to bring AI into the content moderation process on Facebook's platforms. However, not all of them have met with success.

Facebook using artificial intelligence to priorities reported content. Facebook on Tuesday said it has stepped up the use of artificial intelligence (AI) to "prioritise reported content", a move that will help the social media giant take action faster on harmful and violative content. Facebook, which has 1.82 billion daily users globally, has drawn flak in the past for its handling of hate speech on the platform in India, which is among its biggest markets. Facebook Product Manager (Community Integrity) Ryan Barnes said the company is using AI to prioritise reported content, and that this prioritisation is important to help its over 15,000 reviewers. She explained that the prioritisation is important for four reasons -- not all harmful content is equal, some enforcement decisions are complex, people do not always report harmful content and the reports aren't always accurate.

Within intelligent tutoring systems, considerable research has investigated hints, including how to generate data-driven hints, what hint content to present, and when to provide hints for optimal learning outcomes. However, less attention has been paid to how hints are presented. In this paper, we propose a new hint delivery mechanism called "Assertions" for providing unsolicited hints in a data-driven intelligent tutor. Assertions are partially-worked example steps designed to appear within a student workspace, and in the same format as student-derived steps, to show students a possible subgoal leading to the solution. We hypothesized that Assertions can help address the well-known hint avoidance problem. In systems that only provide hints upon request, hint avoidance results in students not receiving hints when they are needed. Our unsolicited Assertions do not seek to improve student help-seeking, but rather seek to ensure students receive the help they need. We contrast Assertions with Messages, text-based, unsolicited hints that appear after student inactivity. Our results show that Assertions significantly increase unsolicited hint usage compared to Messages. Further, they show a significant aptitude-treatment interaction between Assertions and prior proficiency, with Assertions leading students with low prior proficiency to generate shorter (more efficient) posttest solutions faster. We also present a clustering analysis that shows patterns of productive persistence among students with low prior knowledge when the tutor provides unsolicited help in the form of Assertions. Overall, this work provides encouraging evidence that hint presentation can significantly impact how students use them and using Assertions can be an effective way to address help avoidance.

The insurance industry, no stranger to gauging risk, is facing one of its most profound disruptions in decades. Artificial intelligence, machine learning, Internet of Things, blockchain, data analytics and other emerging technologies are enabling many startups to nip at parts of their businesses. Incumbent insurers still have the advantage of institutional knowledge and regulatory expertise, as well as robust cash flows. But they can't sit still. Recognizing the technological winds of change, Reinsurance Group of America (RGA), one of the largest global reinsurers, created RGAX in 2015 to incubate and launch new products and services as its insurer clients seek to maintain their competitive advantages. RGAX CEO Dennis Barnes recently spoke to Knowledge@Wharton about the opportunities and roadblocks to digital transformation. An edited transcript of the conversation follows.