Counterspeech offers direct rebuttals to hateful speech by challenging perpetrators of hate and showing support to targets of abuse. It provides a promising alternative to more contentious measures, such as content moderation and deplatforming, by contributing a greater amount of positive online speech rather than attempting to mitigate harmful content through removal. Advances in the development of large language models mean that the process of producing counterspeech could be made more efficient by automating its generation, which would enable large-scale online campaigns. However, we currently lack a systematic understanding of several important factors relating to the efficacy of counterspeech for hate mitigation, such as which types of counterspeech are most effective, what are the optimal conditions for implementation, and which specific effects of hate it can best ameliorate. This paper aims to fill this gap by systematically reviewing counterspeech research in the social sciences and comparing methodologies and findings with computer science efforts in automatic counterspeech generation. By taking this multi-disciplinary view, we identify promising future directions in both fields.

This paper focuses on an important type of black-box attacks, i.e., transfer-based adversarial attacks, where the adversary generates adversarial examples by a substitute (source) model and utilize them to attack an unseen target model, without knowing its information. Existing methods tend to give unsatisfactory adversarial transferability when the source and target models are from different types of DNN architectures (e.g. ResNet-18 and Swin Transformer). In this paper, we observe that the above phenomenon is induced by the output inconsistency problem. To alleviate this problem while effectively utilizing the existing DNN models, we propose a common knowledge learning (CKL) framework to learn better network weights to generate adversarial examples with better transferability, under fixed network architectures. Specifically, to reduce the model-specific features and obtain better output distributions, we construct a multi-teacher framework, where the knowledge is distilled from different teacher architectures into one student network. By considering that the gradient of input is usually utilized to generated adversarial examples, we impose constraints on the gradients between the student and teacher models, to further alleviate the output inconsistency problem and enhance the adversarial transferability. Extensive experiments demonstrate that our proposed work can significantly improve the adversarial transferability.

Diffusion on graphs is ubiquitous with numerous high-impact applications. In these applications, complete diffusion histories play an essential role in terms of identifying dynamical patterns, reflecting on precaution actions, and forecasting intervention effects. Despite their importance, complete diffusion histories are rarely available and are highly challenging to reconstruct due to ill-posedness, explosive search space, and scarcity of training data. To date, few methods exist for diffusion history reconstruction. They are exclusively based on the maximum likelihood estimation (MLE) formulation and require to know true diffusion parameters. In this paper, we study an even harder problem, namely reconstructing Diffusion history from A single SnapsHot} (DASH), where we seek to reconstruct the history from only the final snapshot without knowing true diffusion parameters. We start with theoretical analyses that reveal a fundamental limitation of the MLE formulation. We prove: (a) estimation error of diffusion parameters is unavoidable due to NP-hardness of diffusion parameter estimation, and (b) the MLE formulation is sensitive to estimation error of diffusion parameters. To overcome the inherent limitation of the MLE formulation, we propose a novel barycenter formulation: finding the barycenter of the posterior distribution of histories, which is provably stable against the estimation error of diffusion parameters. We further develop an effective solver named DIffusion hiTting Times with Optimal proposal (DITTO) by reducing the problem to estimating posterior expected hitting times via the Metropolis--Hastings Markov chain Monte Carlo method (M--H MCMC) and employing an unsupervised graph neural network to learn an optimal proposal to accelerate the convergence of M--H MCMC. We conduct extensive experiments to demonstrate the efficacy of the proposed method.

The events in a narrative are understood as a coherent whole via the underlying states of their participants. Often, these participant states are not explicitly mentioned, instead left to be inferred by the reader. A model that understands narratives should likewise infer these implicit states, and even reason about the impact of changes to these states on the narrative. To facilitate this goal, we introduce a new crowdsourced English-language, Participant States dataset, PASTA. This dataset contains inferable participant states; a counterfactual perturbation to each state; and the changes to the story that would be necessary if the counterfactual were true. We introduce three state-based reasoning tasks that test for the ability to infer when a state is entailed by a story, to revise a story conditioned on a counterfactual state, and to explain the most likely state change given a revised story. Experiments show that today's LLMs can reason about states to some degree, but there is large room for improvement, especially in problems requiring access and ability to reason with diverse types of knowledge (e.g. physical, numerical, factual).

The Federal Trade Commission ( FTC) has proposed a formal ban on fake reviews and testimonials. Companies would also be prohibited from using phony followers and views to inflate their social media metrics if the rule takes effect as it stands. This isn't the first time the agency has trained its sights on fake reviews. In its first such case in 2019, it fined a third-party Amazon seller for paying for fake reviews (Amazon itself has sued phony review providers). Earlier this year, the FTC levied a $600,000 penalty against the owner of a vitamin brand for "review hijacking" on Amazon.

Libya's government denied reports it is responsible for drone attacks that hit an airbase in the east used by mercenaries of the Russian paramilitary group Wagner. The origin of the early Friday attack on the Al-Kharruba airbase, 150km (90 miles) southwest of Benghazi, was unclear but it caused no casualties. Army Chief of Staff General Mohamad al-Haddad denied the Tripoli-based authorities had anything to do with the raid. "None of our aircraft targeted any site in the east," al-Haddad said, according to the Libyan news website Addresslibya. "These reports are aimed at stoking a new war between Libyan brothers and involving Libya in a regional conflict."

Speculation is rampant as US lawmakers continue to voice their own opinions about the explosive claims of an illegal, hidden UFO crash retrieval program made public this month by Air Force and intelligence agency veteran David Grusch. Congressman and Marine veteran Mike Gallagher let loose his own theories on the mystery this Tuesday, suggesting that UFOs might be time-traveling craft piloted by humans from the future, as in the 1984 film'The Terminator.' Appearing on ESPN analyst Pat McAfee's sports talk show, the Wisconsin Republican also floated his hypothesis that the unexplained phenomena'could actually be an ancient civilization that's just been hiding here and is suddenly showing itself.' But Rep. Gallagher also brought the conversation back down to Earth, airing his concerns that the airborne mysteries might prove to be breakthrough aerospace technology mastered by a US foreign adversary. 'I'm probably the most interested in is whether it's adversary technology, particularly from China,' said Gallagher, who is also the chair of the House Committee on the Chinese Communist Party. Whether or not we are alone in the universe, the congressman is not alone among his fellow lawmakers in openly airing his UFO concerns.

In the late 1990s, Recife, on Brazil's northeastern coast, was in decline. Its picturesque historic center, made up of 17th century colonial buildings with Dutch, Portuguese, and French influences, had plunged into neglect, reflecting a deep economic crisis worsened by deindustrialization. Many young people were fleeing the city for opportunities in the commercial centers of São Paulo and Rio de Janeiro, or heading overseas. In 2000, a group of businesspeople, government officials, and academics came up with a vision to regenerate Recife's historic center by building a new technology district. With 33 million reais ($6.8 million) raised from the privatization of the local electricity company, they created Porto Digital, a nonprofit organization with the mission of turning Recife into a hub for technology and the creative industries.

Artificial intelligence is either going to save humanity or finish it off, depending on who you speak to. Either way, every week there are new developments and breakthroughs. The US company behind the ChatGPT chatbot, OpenAI, has announced that its first international office will be in London. The move is a boost for the UK prime minister, Rishi Sunak, who has described the AI race as one of the "greatest opportunities" for the country's tech industry. OpenAI said it chose the UK capital because of its "rich culture and exceptional talent pool".

Center for A.I. Safety Director Dan Hendrycks explains concerns about how the rapid growth of artificial intelligence could impact society. House lawmakers are urging federal agencies to quickly and aggressively adopt artificial intelligence technology, at a time when the push from civil rights and industry groups for new AI regulations is still waiting to get off the ground. The House Appropriations Committee, led by Rep. Kay Granger, R-Texas, released several spending bills this week that encourage the government to incorporate AI into everything from national security functions to routine office work to the detection of pests and diseases in crops. Several of those priorities are not just encouraged but would get millions of dollars in new funding under the legislation still being considered by the committee. And while comprehensive AI regulations are likely still months away and are unlikely to be developed this year, lawmakers seem keen on making sure the government is deploying AI where it can. The bills are backed by the GOP majority, and Rep. Don Beyer, D-Va., the vice chair of the Congressional Artificial Intelligence Caucus, said agencies shouldn't have to wait to start using AI.