Fact-checking requires retrieving evidence related to a claim under investigation. The task can be formulated as question generation based on a claim, followed by question answering. However, recent question generation approaches assume that the answer is known and typically contained in a passage given as input, whereas such passages are what is being sought when verifying a claim. In this paper, we present {\it Varifocal}, a method that generates questions based on different focal points within a given claim, i.e.\ different spans of the claim and its metadata, such as its source and date. Our method outperforms previous work on a fact-checking question generation dataset on a wide range of automatic evaluation metrics. These results are corroborated by our manual evaluation, which indicates that our method generates more relevant and informative questions. We further demonstrate the potential of focal points in generating sets of clarification questions for product descriptions.

While recent work has convincingly showed that sequence-to-sequence models struggle to generalize to new compositions (termed compositional generalization), little is known on what makes compositional generalization hard on a particular test instance. In this work, we investigate what are the factors that make generalization to certain test instances challenging. We first substantiate that indeed some examples are more difficult than others by showing that different models consistently fail or succeed on the same test instances. Then, we propose a criterion for the difficulty of an example: a test instance is hard if it contains a local structure that was not observed at training time. We formulate a simple decision rule based on this criterion and empirically show it predicts instance-level generalization well across 5 different semantic parsing datasets, substantially better than alternative decision rules. Last, we show local structures can be leveraged for creating difficult adversarial compositional splits and also to improve compositional generalization under limited training budgets by strategically selecting examples for the training set.

We propose a new problem called coordinated topic modeling that imitates human behavior while describing a text corpus. It considers a set of well-defined topics like the axes of a semantic space with a reference representation. It then uses the axes to model a corpus for easily understandable representation. This new task helps represent a corpus more interpretably by reusing existing knowledge and benefits the corpora comparison task. We design ECTM, an embedding-based coordinated topic model that effectively uses the reference representation to capture the target corpus-specific aspects while maintaining each topic's global semantics. In ECTM, we introduce the topic- and document-level supervision with a self-training mechanism to solve the problem. Finally, extensive experiments on multiple domains show the superiority of our model over other baselines.

Maximum likelihood estimation (MLE) is the predominant algorithm for training text generation models. This paradigm relies on direct supervision examples, which is not applicable to many emerging applications, such as generating adversarial attacks or generating prompts to control language models. Reinforcement learning (RL) on the other hand offers a more flexible solution by allowing users to plug in arbitrary task metrics as reward. Yet previous RL algorithms for text generation, such as policy gradient (on-policy RL) and Q-learning (off-policy RL), are often notoriously inefficient or unstable to train due to the large sequence space and the sparse reward received only at the end of sequences. In this paper, we introduce a new RL formulation for text generation from the soft Q-learning (SQL) perspective. It enables us to draw from the latest RL advances, such as path consistency learning, to combine the best of on-/off-policy updates, and learn effectively from sparse reward. We apply the approach to a wide range of novel text generation tasks, including learning from noisy/negative examples, adversarial attacks, and prompt generation. Experiments show our approach consistently outperforms both task-specialized algorithms and the previous RL methods.

Offline reinforcement learning (RL) provides a promising direction to exploit massive amount of offline data for complex decision-making tasks. Due to the distribution shift issue, current offline RL algorithms are generally designed to be conservative in value estimation and action selection. However, such conservatism can impair the robustness of learned policies when encountering observation deviation under realistic conditions, such as sensor errors and adversarial attacks. To trade off robustness and conservatism, we propose Robust Offline Reinforcement Learning (RORL) with a novel conservative smoothing technique. In RORL, we explicitly introduce regularization on the policy and the value function for states near the dataset, as well as additional conservative value estimation on these states. Theoretically, we show RORL enjoys a tighter suboptimality bound than recent theoretical results in linear MDPs. We demonstrate that RORL can achieve state-of-the-art performance on the general offline RL benchmark and is considerably robust to adversarial observation perturbations.

Fox News national security correspondent Jennifer Griffin provides insight on responding to drone attacks in Ukraine on'America Reports.' "Star Wars" actor Mark Hamill donated 500 drones to assist Ukraine in its efforts to push back Russian forces. "Very simply: Ukraine needs drones," Hamill told host Joe Mathieu during an appearance on Bloomberg Radio's "Sound On." "They define war outcomes, they protect their land, their people, they monitor the border, they're eyes in the sky," Hamill said, adding that he was "honored" to support the mission, which would include raising funds to support Ukraine's UNITED24 "Army of Drones" project. Ukrainian President Volodymyr Zelenskyy announced the program's launch last year, including Hamill's role as an ambassador for the program. So far, donations through UNITED24 have reached more than $210 million. CONOR KENNEDY, RFK'S GRANDSON AND TAYLOR SWIFT'S EX, REVEALS UKRAINE SERVICE: 'I WAS … WILLING TO DIE' "I was really shocked because they give me these updates at least two or three times a week about what's going on, and they said they've received over 500 drones since I started this," the actor revealed.

NASA has recruited the best of the best to unravel the mysteries of unexplained natural phenomena seen in the skies with the hopes of solving hundreds of sightings. Astronaut Scott Kelly, who is famously known for spending nearly a year in space, is one of 16 team members who will comb through unclassified data of UFOs to help the space agency'craft scientific conclusions' about what is actually occurring. The independent study is set to launch Monday and will carry on for the next nine months - the team is expected to share their findings with the public in mid-2023. The program, announced in June, follows the first US congressional hearings in 50 years that revealed videos showing 144 'unidentified aerial phenomena' seen by military personnel since 2004. The team, which includes scientists, data and artificial intelligence specialists and aerospace safety experts, will identify how data gathered by civilian government entities, commercial data, and data from other sources can potentially be analyzed to shed light on unidentified aerial phenomena (UAP).

Artificial intelligence (AI) and data science (DS) centers are becoming ubiquitous in academic institutions around the globe. These centers serve to focus research efforts and bring together large teams to address important problems. AI centers in more mature research ecosystems tend to be multi-institutional, such as the Alan Turing Institute in the U.K. with 13 academic partners12 and Mila in Montreal with four academic partners and numerous industry partners.8 Often such centers are also focused on a specific theme, such as the 18 AI institutes funded by NSF.10 In contrast, the centers in India tend to be contained in only one institute--this facilitates the institute to identify AI/DS as a growth area and an area of interest to the Institute.

Deep learning (DL) models have enjoyed tremendous success across application domains within the broader umbrella of artificial intelligence (AI) technologies. However, their "black-box" nature, coupled with their extensive use across application sectors--including safety-critical and risk-sensitive ones such as healthcare, finance, aerospace, law enforcement, and governance--has elicited an increasing need for explainability, interpretability, and transparency of decision-making in these models.11,14,18,24 With the recent progression of legal and policy frameworks that mandate explaining decisions made by AI-driven systems (for example, the European Union's GDPR Article 15(1)(h) and the Algorithmic Accountability Act of 2019 in the U.S.), explainability has become a cornerstone of responsible AI use and deployment. In the Indian context, NITI Aayog recently released a two-part strategy document on envisioning and operationalizing Responsible AI in India,15,16 which puts significant emphasis on the explainability and transparency of AI models. Explainability of DL models lies at the human-machine interface, and different users may expect different explanations in different contexts.

Microprocessors and microcontrollers form the core of electronic systems. Unfortunately, almost all the microprocessors and microcontrollers are imported, currently. IPs and patents with strict licensing terms protect their designs. Realizing the limitations of the processor industry, the SHAKTI Processor Program11 started as an academic initiative back in 2014. SHAKTI is an open source processor4 initiative by the Pratap Subramaniam–Center for Digital Intelligence and Secure Hardware Architecture (PC-CDISHA)–Reconfigurable Intelligent Systems Engineering (RISE) group, IIT-Madras.