link1s.site

Stanford AI project team apologizes for plagiarizing Chinese model

An artificial intelligence (AI) team at Stanford University apologized for plagiarizing a large language model (LLM) from a Chinese AI company, which became a trending topic on the Chinese social media platforms, where it sparked concern among netizens on Tuesday.

We apologize to the authors of MiniCPM [the AI model developed by a Chinese company] for any inconvenience that we caused for not doing the full diligence to verify and peer review the novelty of this work, the multimodal AI model Llama3-V's developers wrote in a post on social platform X.

The apology came after the team from Stanford University announced Llama3-V on May 29, claiming it had comparable performance to GPT4-V and other models with the capability to train for less than $500.

According to media reports, the announcement published by one of the team members quickly received more than 300,000 views.

However, some netizens from X found and listed evidence of how the Llama3-V project code was reformatted and similar to MiniCPM-Llama3-V 2.5, an LLM developed by a Chinese technology company, ModelBest, and Tsinghua University.

Two team members, Aksh Garg and Siddharth Sharma, reposted a netizen's query and apologized on Monday, while claiming that their role was to promote the model on Medium and X (formerly Twitter), and that they had been unable to contact the member who wrote the code for the project.

They looked at recent papers to validate the novelty of the work but had not been informed of or were aware of any of the work by Open Lab for Big Model Base, which was founded by the Natural Language Processing Lab at Tsinghua University and ModelBest, according to their responses. They noted that they have taken all references to Llama3-V down in respect to the original work.

In response, Liu Zhiyuan, chief scientist at ModelBest, spoke out on the Chinese social media platform Zhihu, saying that the Llama3-V team failed to comply with open-source protocols for respecting and honoring the achievements of previous researchers, thus seriously undermining the cornerstone of open-source sharing.

According to a screenshot leaked online, Li Dahai, CEO of ModelBest, also made a post on his WeChat moment, saying that the two models were verified to have highly similarity in terms of providing answers and even the same errors, and that some relevant data had not yet been released to the public.

He said the team hopes that their work will receive more attention and recognition, but not in this way. He also called for an open, cooperative and trusting community environment.

Director of the Stanford Artificial Intelligence Laboratory Christopher Manning also responded to Garg's explanation on Sunday, commenting "How not to own your mistakes!" on X.

As the incident became a trending topic on Sina Weibo, Chinese netizens commented that academic research should be factual, but the incident also proves that the technology development in China is progressing.

Global Times

Adult Film Star Jesse Jane's Cause of Death Revealed
New details about Jesse Jane’s death have emerged. More than six months after the former adult film star was found dead alongside her boyfriend Brett Hasenmueller in her Moore, Oklahoma, home from a suspected overdose, authorities confirmed what led to the tragedy. Jane and Hasenmueller reportedly died of an accidental fentanyl and cocaine overdose according to TMZ, citing the Oklahoma City Medical Examiner’s Office. E! News has reached out to the Examiner’s Office for comment but had not yet heard back. In late January, local outlet KFOR reported that Moore Police found Jane and Hasemueller’s bodies while performing a wellness check after the couple had not been heard from in several days. Moore PD's Lt. Francisco Franco told The New York Times at the time that it was believed Jane died from a drug overdose, but that both deaths remained under investigation. Jane, who was 43 at the time of her passing, acted in a number of adult films starting in 2003 before transitioning to other projects, including an appearance in Baywatch: Hawaiian Wedding and 2004's Starsky and Hutch, as well as season two of Entourage. Pretty Pastel Please, YouTuber, Instagram By 2007, Jane had officially retired from the adult film industry, pivoting into making her own line of sex toys. At the time of her death, BSG Public Relations President Brian Gross shared in a statement to E! News, "Jesse Jane was a vivacious person who had an absolute and ultimate love for life. During her time in the adult industry, of which I was able to spend wonderful moments with her, she was an incredible professional who cheered everyone on and brought sunshine to every film set she worked on." "There is not one person in the adult industry who didn't spend time with her, whether onset or in a social setting, that she didn't make smile, laugh or both. She would light up a room as soon as she walked in," he continued, "I will personally miss her very much for the reasons above. Her smile was everything."
NASA plays 'blame-shifting' game with China as lunar soil research set to start
The returner of the Chang'e-6 lunar probe is opened during a ceremony at the China Academy of Space Technology under the China Aerospace Science and Technology Corporation in Beijing, capital of China, June 26, 2024. The returner of the Chang'e-6 lunar probe was opened at a ceremony in Beijing on Wednesday afternoon. During the ceremony at the China Academy of Space Technology under the China Aerospace Science and Technology Corporation, researchers opened the returner and examined key technical indicators. Photo: Xinhua As the US space industry recently faced yet more delays and stagnation with key components including manned spacecraft and space suits "going wrong," NASA has once again resorted to its "sour grapes" rhetoric upon seeing China's successful retrieval of fresh lunar soils from the far side of the moon, by claiming that China did not directly invite its scientists to participate in the lunar soil research. This behavior is a typical blame-shifting trick, Chinese experts said, noting it is clear to all that it is the US' own laws, not China, that are restricting space cooperation between the two sides. Instead of deceiving themselves by distorting the truth, the US should face up to its own problem of overall weakening engineering capability and the lack of long-term planning in its space industry. After the Chang'e-6 samples, weighing nearly 2 kilograms, were safely transported to a special laboratory for further study on Friday, NASA spokesperson Faith McKie told media that while China worked with the European Space Agency, France, Italy and Pakistan on this mission, "NASA wasn't invited to take part in the moon probe." NASA also didn't get "any direct invitation" to study China's moon rocks, after it welcomed all scientists from around the world to apply to study them, McKie told NatSec Daily. Responding to the remarks, Chinese Foreign Ministry spokesperson Mao Ning told the Global Times on Monday that China is open to having space exchanges with the US, and we also welcome countries around the world to take part in the study of lunar samples. "However, the US side seems to have forgotten to mention its domestic legislation such as the Wolf Amendment. The real question is whether US scientists and institutions are allowed by their own government to participate in cooperation with China," Mao said. "The existence of the Wolf Amendment has basically shut the door to space collaboration between the two countries," Wang Yanan, chief editor of Beijing-based Aerospace Knowledge magazine, told the Global Times on Monday. Even if research institutions of the US have the willingness to work with China on opportunities such as lunar sample research, institutions there must obtain special approval from the US Congress due to the presence of this amendment, Wang explained. Currently, no such "green light" is in sight from the Congress. Furthermore, China's collaboration with international partners is based on equality and mutual benefit, leveraging their respective scientific resources, facilities, and expertise. However, the US only wants what it doesn't have, and its engagement with China would be advantageous only to itself, Wang noted. NASA has found itself embroiled in a number of thorny issues recently, with the latest being Boeing's Starliner manned spaceship experiencing both helium leaks and thruster issues during a June 6 docking with the International Space Station (ISS), which led to an indefinite delay for its crew's return to Earth, despite NASA's insistence that they are not "stranded" in space. The return of the Starliner capsule, while has already been delayed by two weeks, will be put on hold "well into the summer" pending results of new thruster tests, which are scheduled to start Tuesday and will take approximately two weeks or even more, per NASA officials. Previously on June 24, NASA cancelled a spacewalk on the ISS following a "serious situation," when one of the spacesuits experienced coolant leak in the hatch. While being broadcast on a livestream, the astronauts reported "literally water everywhere" as they were preparing for the extravehicular activity, space.com reported. The report said that this is the second time this particular spacewalk was postponed, after a June 13 attempt with a different astronaut group was pushed back due to a "spacesuit discomfort." The recurring issues with the spacesuits are due to their much-extended service lifespan, media reported, as the puffy white ones US astronauts currently wear were designed more than 40 years ago. Despite the pressing need to replace them, NASA announced recently that it is abandoning a plan to develop next-generation spacesuits, which had been committed to be delivered by 2026, CNN reported on Thursday. One of the root causes for such problems is that the US has developed many large technology conglomerates, which for a long time have benefited significantly from government orders and industry monopolies. Consequently, in many complex engineering fields, the level of attention given is greatly insufficient, Wang noted. It also reflected the US' lack of long-term strategic planning for its manned space program. For instance, the ageing spacesuits should have been replaced a decade ago to ensure that operational suits remain in usable condition. Failure to address this issue results in a hindrance to the space station's necessary maintenance tasks and even poses life-threatening risks to astronauts in emergency situations, experts said. The issues with Boeing's spacecraft and the spacesuits are not isolated problems, but reflected a systemic issue in the US space industry - the overall weakening of engineering capabilities, they noted.
The largest password leak in history exposes nearly 10 billion credentials
The largest collection of stolen passwords ever has been leaked to a notorious crime marketplace, according to cybersecurity researchers at Cybernews. This leak, dubbed RockYou2024 by its original poster “ObamaCare,” holds a file containing nearly 10 billion unique plaintext passwords. Allegedly gathered from a series of data breaches and hacks accumulated over several years, the passwords were posted on July 4th and hailed as the most extensive collection of stolen and leaked credentials ever seen on the forum. “In its essence, the RockYou2024 leak is a compilation of real-world passwords used by individuals all over the world,” the researchers told Cybernews. “Revealing that many passwords for threat actors substantially heightens the risk of credential stuffing attacks.” Credential stuffing attacks are among the most common methods criminals, ransomware affiliates, and state-sponsored hackers use to access services and systems. Threat actors could exploit the RockYou2024 password collection to conduct brute-force attacks against any unprotected system and “gain unauthorized access to various online accounts used by individuals whose passwords are included in the dataset,” the research team said. This could affect online services, cameras and hardware This could affect various targets, from online services to internet-facing cameras and industrial hardware. “Moreover, combined with other leaked databases on hacker forums and marketplaces, which, for example, contain user email addresses and other credentials, RockYou2024 can contribute to a cascade of data breaches, financial frauds, and identity thefts,” the team concluded. However, despite the seriousness of the data leak, it is important to note that RockYou2024 is primarily a compilation of previous password leaks, estimated to contain entries from a total of 4,000 massive databases of stolen credentials, covering at least two decades. This new file notably includes an earlier credentials database known as RockYou2021, which featured 8.4 billion passwords. RockYou2024 added approximately 1.5 billion passwords to the collection, spanning from 2021 through 2024, which, though a massive figure, is only a fraction of the reported 9,948,575,739 passwords in the leak. Thus, users who have changed their passwords since 2021 may not have to panic about a potential breach of their information. That said, the research team at Cybernews stressed the importance of maintaining data security. In response to the leak, they recommend immediately changing the passwords for any accounts associated with the leaked credentials, ensuring each password is strong and unique and not reused across different platforms. Additionally, they advised enabling multi-factor authentication (MFA), which requires an extra form of verification beyond the password, wherever possible, to strengthen cyber security. Lastly, tech users should utilize password manager software, which securely generates and stores complex passwords, mitigating the risk of password reuse across multiple accounts.
Argentina's government reform bill officially takes effect: granting the president special powers in areas such as administration
On the 8th, the Argentine government promulgated the "Foundations and Starting Points for Argentine Freedom" comprehensive bill and a package of fiscal measures, marking the official entry into force of the government reform bill. According to the official gazette of the Argentine government, Argentine President Milley, Chief Cabinet Minister Guillermo Francos and Economy Minister Luis Caputo jointly signed Decrees No. 592 and No. 593 to promulgate these two new reform measures. The comprehensive bill declared Argentina to enter a one-year public emergency in the administrative, economic, financial and energy fields, and granted the president special powers in these fields. It also includes the relaxation of economic regulations, labor reforms and the implementation of a large-scale investment incentive system. The package of fiscal measures involves anti-money laundering, tax deferral, tariffs, re-imposition of high-salary income tax and reduction of personal property taxes. On June 28, after six months of negotiations, the two reform bills were finally passed by the Argentine Congress.
Stanford AI project team apologizes for plagiarizing Chinese model
An artificial intelligence (AI) team at Stanford University apologized for plagiarizing a large language model (LLM) from a Chinese AI company, which became a trending topic on the Chinese social media platforms, where it sparked concern among netizens on Tuesday. We apologize to the authors of MiniCPM [the AI model developed by a Chinese company] for any inconvenience that we caused for not doing the full diligence to verify and peer review the novelty of this work, the multimodal AI model Llama3-V's developers wrote in a post on social platform X. The apology came after the team from Stanford University announced Llama3-V on May 29, claiming it had comparable performance to GPT4-V and other models with the capability to train for less than $500. According to media reports, the announcement published by one of the team members quickly received more than 300,000 views. However, some netizens from X found and listed evidence of how the Llama3-V project code was reformatted and similar to MiniCPM-Llama3-V 2.5, an LLM developed by a Chinese technology company, ModelBest, and Tsinghua University. Two team members, Aksh Garg and Siddharth Sharma, reposted a netizen's query and apologized on Monday, while claiming that their role was to promote the model on Medium and X (formerly Twitter), and that they had been unable to contact the member who wrote the code for the project. They looked at recent papers to validate the novelty of the work but had not been informed of or were aware of any of the work by Open Lab for Big Model Base, which was founded by the Natural Language Processing Lab at Tsinghua University and ModelBest, according to their responses. They noted that they have taken all references to Llama3-V down in respect to the original work. In response, Liu Zhiyuan, chief scientist at ModelBest, spoke out on the Chinese social media platform Zhihu, saying that the Llama3-V team failed to comply with open-source protocols for respecting and honoring the achievements of previous researchers, thus seriously undermining the cornerstone of open-source sharing. According to a screenshot leaked online, Li Dahai, CEO of ModelBest, also made a post on his WeChat moment, saying that the two models were verified to have highly similarity in terms of providing answers and even the same errors, and that some relevant data had not yet been released to the public. He said the team hopes that their work will receive more attention and recognition, but not in this way. He also called for an open, cooperative and trusting community environment. Director of the Stanford Artificial Intelligence Laboratory Christopher Manning also responded to Garg's explanation on Sunday, commenting "How not to own your mistakes!" on X. As the incident became a trending topic on Sina Weibo, Chinese netizens commented that academic research should be factual, but the incident also proves that the technology development in China is progressing. Global Times