Large neural networks can now generate jokes, but do they really "understand" humor? We challenge AI models with three tasks derived from the New Yorker Cartoon Caption Contest: matching a joke to a cartoon, identifying a winning caption, and explaining why a winning caption is funny. These tasks encapsulate progressively more sophisticated aspects of "understanding" a cartoon; key elements are the complex, often surprising relationships between images and captions and the frequent inclusion of indirect and playful allusions to human experience and culture. We investigate both multimodal and language-only models: the former are challenged with the cartoon images directly, while the latter are given multifaceted descriptions of the visual scene to simulate human-level visual understanding. We find that both types of models struggle at all three tasks. For example, our best multimodal models fall 30 accuracy points behind human performance on the matching task, and, even when provided ground-truth visual scene descriptors, human-authored explanations are preferred head-to-head over the best machine-authored ones (few-shot GPT-4) in more than 2/3 of cases. We release models, code, leaderboard, and corpus, which includes newly-gathered annotations describing the image's locations/entities, what's unusual in the scene, and an explanation of the joke.

We propose end-to-end multimodal fact-checking and explanation generation, where the input is a claim and a large collection of web sources, including articles, images, videos, and tweets, and the goal is to assess the truthfulness of the claim by retrieving relevant evidence and predicting a truthfulness label (e.g., support, refute or not enough information), and to generate a statement to summarize and explain the reasoning and ruling process. To support this research, we construct Mocheg, a large-scale dataset consisting of 15,601 claims where each claim is annotated with a truthfulness label and a ruling statement, and 33,880 textual paragraphs and 12,112 images in total as evidence. To establish baseline performances on Mocheg, we experiment with several state-of-the-art neural architectures on the three pipelined subtasks: multimodal evidence retrieval, claim verification, and explanation generation, and demonstrate that the performance of the state-of-the-art end-to-end multimodal fact-checking does not provide satisfactory outcomes. To the best of our knowledge, we are the first to build the benchmark dataset and solutions for end-to-end multimodal fact-checking and explanation generation. The dataset, source code and model checkpoints are available at https://github.com/VT-NLP/Mocheg.

Former U.S. Capitol Police Chief Steven Sund joins'Hannity' to sound off on bureaucratic failures leading up to riot. A January 6 defendant who was allegedly found with weapons and materials to make an explosive device just blocks from former President Barack Obama's home in Washington, D.C. last week threatened to blow up a van at a government facility and attempted to threaten a congressman, federal prosecutors said Tuesday. Taylor Taranto, 37, a Washington state resident, was wanted for allegedly participating in the Jan. 6 riot when he was taken into custody by Secret Service agents on June 29 in the Kalorama neighborhood of Washington D.C., according to court documents. The arrest came a day after he live-streamed himself on his public YouTube channel when he said he had a detonator and threatened to blow up his "self-driving" vehicle at the National Institute of Standards and Technology, which is headquartered in Gaithersburg, Maryland, prosecutors said. Taylor Taranto, 37, an accused Jan. 6 rioter, was arrested last week with weapons near former President Barack Obama's home in Washington D.C., prosecutors said.

Robocall scams steal tens of billions of dollars every year in the U.S. Frustrated by the constant calls, this man created an AI to waste scammers' time. After a telemarketer swore at his son nearly 10 years ago, Roger Anderson set out to cause robocallers, telephone fraudsters and scammers as many problems as possible. His solution was to create an artificial intelligence call-deflection system that puts incoming spam calls on the line with a ChatGPT-powered personality, keeping fraudsters talking in circles as they attempt to extract financial information from a chatbot. "This is really an incredibly silly idea," Anderson, who founded the Jolly Roger Telephone Company, told Fox News. "But in a way, I think it's the only way to combat unsolicited telemarketing because we just can't seem to get a track on stopping all of these unwanted calls."

OpenAI is forming a dedicated team to manage the risks of superintelligent artificial intelligence. A superintelligence is a hypothetical AI model that is smarter than even the most gifted and intelligent human, and excels at multiple areas of expertise instead of one domain like some previous generation models. OpenAI believes such a model could arrive before the end of the decade. "Superintelligence will be the most impactful technology humanity has ever invented, and could help us solve many of the world's most important problems," the non-profit said. "But the vast power of superintelligence could also be very dangerous, and could lead to the disempowerment of humanity or even human extinction."

Doctors believe Artificial Intelligence is now saving lives, after a major advancement in breast cancer screenings. A.I. is detecting early signs of the disease, in some cases years before doctors would find the cancer on a traditional scan. Department of Veterans Affairs researchers at five medical centers are working together to develop an artificial intelligence algorithm that can predict aggressive prostate cancer. The new research study began July 1, expanding to 14 sites. It will analyze data from more than 5,000 veterans who were diagnosed with high-risk prostate cancer and have undergone initial treatment.

PactumAI co-founder and CEO Martin Rand explains how workers can use artificial intelligence to enhance their careers and positions. 'EVOLUTION' OF'TRADITION': Harvard University will begin to incorporate the use of chatbots to teach it's coding class this fall. Professor Malan states the hopes are to create a one-to-one ratio for teachers and students by providing software-based tools. DOOMSDAY CONTROL: An artificial intelligence system may have the ability to launch nukes on its own. U.S. lawmakers are proposing bipartisan legislation to keep America's nuclear arsenal under human control.

Hall of Fame tennis coach Rick Macci weighs in on how fans will react to a computer commentator instead of a human one on'Fox & Friends.' The United Nations Security Council is holding its first-ever meeting on the potential risks artificial intelligence poses to the maintenance of international peace and security. Organized by the United Kingdom, U.K. Ambassador Barbara Woodward announced the July 18 gathering on Monday. The talks will include remarks from experts in the emergent field, as well as input from U.N. Secretary-General Antonio Guterres. Last month, he warned that alarm bells over the most advanced form of AI are "deafening."

Center for A.I. Safety Director Dan Hendrycks explains concerns about how the rapid growth of artificial intelligence could impact society. The possibility that an artificial intelligence system might launch a nuclear attack on its own has prompted House lawmakers to propose legislative language that would ensure America's nuclear arsenal remains under human control. Rep. Ted Lieu, D-Calif., has proposed a bipartisan amendment to the 2024 defense policy bill that requires the Pentagon to put in place a system that ensures "meaningful human control is required to launch any nuclear weapon." It defines human control by saying people must have the final say in selecting and engaging targets, including when, where and how they are hit with a nuclear weapon. It is a concept that senior military leaders say they are already following.

Moscow – Moscow has said that Ukrainian drone attacks on Russian territory would "not be possible" without U.S. and NATO help, escalating its rhetoric after reporting it had downed five drones near the capital on Tuesday. Ukraine meanwhile accused Russia of planning "dangerous provocations" at the Moscow-controlled Zaporizhzhia nuclear power plant, while Russia in turn claimed Kyiv was planning to attack the facility -- Europe's largest. Moscow said the West had enabled Ukraine to carry out the drone attacks, after earlier condemning what it called a "terrorist act." This could be due to a conflict with your ad-blocking or security software. Please add japantimes.co.jp and piano.io to your list of allowed sites.