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

iPhone 16 Pro leak just confirmed a huge camera upgrade
The tetraprism lens with 5x optical zoom currently exclusive to the iPhone 15 Pro Max could be headed to both the iPhone 16 Pro and iPhone 16 Pro Max, narrowing the gap between Apple's premium flagships. That's according to a new report from analyst Ming-Chi Kuo, who cites a recent earnings call with Apple lens supplier Largan. In the call, a spokesperson from Largan said "some flagship specifications will be extended to other models" in the second half of 2024, presumably in reference to the upcoming iPhone Pro models. "Apple is Largan’s largest customer, and Largan is also Apple’s largest lens supplier," Kuo said. "Therefore, the quote likely refers to the fact that the new iPhone 16 Pro and Pro Max will have a tetraprism camera in 2H24 (while only the iPhone 15 Pro Max had this camera in 2H23).” The report goes on to say that the tetraprism camera for the iPhone 16 Pro series won't be all that different from the one in the iPhone 15 Pro Max. While the lack of an upgrade is disappointing, it's not necessarily a bad thing as these kinds of lenses are already top-of-the-line. They represent a major increase over prior models’ zoom capabilities, and they're capable of offering more depth while still fitting into super-slim smartphones. That being said, Apple does appear to be revamping the main camera and ultra-wide camera on the iPhone 16 Pro Max. Evidence continues to mount that both iPhone 16 Pro models will share the same 5x optical zoom camera. Earlier this week, DigitTimes in Asia (via 9to5Mac) reported that Apple is set to ramp up orders for tetraprism lenses as it expands their use in its upcoming iPhone series. Industry sources told the outlet that Largan and Genius Electronic Optical were tapped as the primary suppliers. Apple would be wise to streamline its Pro-level iPhones with the same camera setup; then all customers have to consider with their choice of a new iPhone is the size and price. Of course, this should all be taken with a grain of sand for now until we hear more from Apple. It's still a while yet before Apple's usual September time window for iPhone launches. In the meantime, be sure to check out all the rumors so far in our iPhone 16, iPhone 16 Pro and iPhone 16 Pro Max hubs.
ChatGPT: Explained to Kids(How ChatGPT works)
Chat means chat, and GPT is the acronym for Gene Rate Pre trained Transformer. Genrative means generation, and its function is to create or produce something new; Pre trained refers to a model of artificial intelligence that is learned from a large amount of textual materials, while Transformer refers to a model of artificial intelligence. Don't worry about T, just focus on the words G and P. We mainly use its Generative function to generate various types of content; But we need to know why it can produce various types of content, and the reason lies in P. Only by learning a large amount of content can we proceed with reproduction. And this kind of learning actually has limitations, which is very natural. For example, if you have learned a lot of knowledge since childhood, can you guarantee that your answer to a question is completely correct? Almost impossible, firstly due to the limitations of knowledge, ChatGPT is no exception, as it is impossible to master all knowledge; The second is the accuracy of knowledge, how to ensure that all knowledge is accurate and error free; The third aspect is the complexity of knowledge, where the same concept is manifested differently in different contexts, making it difficult for even humans to grasp it perfectly, let alone AI. So when we use ChatGPT, we also need to monitor the accuracy of the output content of ChatGPT. It is likely not a problem, but if you want to use it on critical issues, you will need to manually review it again. And now ChatGPT has actually been upgraded twice, one is GPT4 with more accurate answering ability, and the other is the recent GPT Turbo. The current ChatGPT is a large model called multimodality, which differs from the first generation in that it can not only receive and output text, but also other types of input, such as images, documents, videos, etc. The output is also more diverse. In addition to text, it can also output images or files, and so on.
Former British PM Sunak appoints Conservative Party shadow cabinet
On July 8, local time, former British Prime Minister Sunak announced the appointment of the Conservative Party Shadow Cabinet, which is the first shadow cabinet of the Conservative Party in 14 years. Several former British cabinet members during Sunak's tenure as prime minister were appointed to the Conservative Party Shadow Cabinet, including James Cleverly as Shadow Home Secretary and Jeremy Hunt as Shadow Chancellor of the Exchequer. But former Foreign Secretary Cameron was not appointed as Shadow Foreign Secretary. In addition, the new leader of the Conservative Party will be elected as early as this week. On July 4, the UK held a parliamentary election. The counting results showed that the British Labour Party won more than half of the seats and won an overwhelming victory; the Conservative Party suffered a disastrous defeat, ending its 14-year continuous rule.
OpenAI's internal AI details stolen in 2023 breach, NYT reports
July 4 (Reuters) - A hacker gained access to the internal messaging systems at OpenAI last year and stole details about the design of the company's artificial intelligence technologies, the New York Times reported, opens new tab on Thursday. The hacker lifted details from discussions in an online forum where employees talked about OpenAI's latest technologies, the report said, citing two people familiar with the incident. However, they did not get into the systems where OpenAI, the firm behind chatbot sensation ChatGPT, houses and builds its AI, the report added. OpenAI executives informed both employees at an all-hands meeting in April last year and the company's board about the breach, according to the report, but executives decided not to share the news publicly as no information about customers or partners had been stolen. OpenAI executives did not consider the incident a national security threat, believing the hacker was a private individual with no known ties to a foreign government, the report said. The San Francisco-based company did not inform the federal law enforcement agencies about the breach, it added. OpenAI in May said it had disrupted five covert influence operations that sought to use its AI models for "deceptive activity" across the internet, the latest to stir safety concerns about the potential misuse of the technology. The Biden administration was poised to open up a new front in its effort to safeguard the U.S. AI technology from China and Russia with preliminary plans to place guardrails around the most advanced AI Models including ChatGPT, Reuters earlier reported, citing sources.
US foreign policy is advanced smartphone with weak battery
A couple of days ago, a Quad summit meeting in Sydney scheduled for May 24 was abruptly canceled. The US president had to pull out of his long-anticipated trip to Australia and Papua New Guinea. Instead, the heads of the four Quad member states got together on the margins of the G7 Summit in Hiroshima on May 20. The main reason for the change of plans was the continuous struggle between the White House and Republicans on the Hill over the national debt ceiling. If no compromise is reached, the US federal government might fail to meet its financial commitments already in June; such a technical default would have multiple negative repercussions for the US, as well as for the global economy and finance at large. Let us hope that a compromise between the two branches of US power will be found and that the ceiling of the national debt will be raised once again. However, this rather awkward last-minute cancellation of the Quad summit reflects a fundamental US problem - a growing imbalance between the US geopolitical ambitions and the fragility of the national financial foundation to serve these ambitions. The Biden administration appears to be fully committed to bringing humankind back to the unipolar world that existed right after the end of the Cold War some 30 years ago, but the White House no longer has enough resources at its disposal to sustain such an undertaking. As they say in America: You cannot not have champagne on a beer budget. The growing gap between the ends that the US seeks in international relations and the means that it has available is particularly striking in the case of the so-called dual containment policy that Washington now pursues toward Russia and China. Even half a century ago, when the US was much stronger in relative terms than it is today, the Nixon administration realized that containing both Moscow and Beijing simultaneously was not a good idea: "Dual containment" would imply prohibitively high economic costs for the US and would result in too many unpredictable political risks. The Nixon administration decided to focus on containing the Soviet Union as the most important US strategic adversary of the time. This is why Henry Kissinger flew to Beijing in July 1971 to arrange the first US-China summit in February 1972 leading to a subsequent rapid rapprochement between the two nations. In the early days of the Biden administration, it seemed that the White House was once again trying to avoid the unattractive "dual containment" option. The White House rushed to extend the New START in January 2021 and held an early US-Russia summit meeting five months later in Geneva. At that point many analysts predicted that Biden would play Henry Kissinger in reverse - that is he would try to peace with the relatively weaker opponent (Moscow) in order to focus on containing the stronger one (Beijing). However, after the beginning of the Russia-Ukraine conflict, it became clear that no accommodation with the Kremlin was on Biden's mind any longer. Still, having decided to take a hard-line stance toward Moscow and to lead a broad Western coalition in providing military and economic assistance to Kiev, Washington has not opted for a more accommodative or at least a more flexible policy toward Beijing. On the contrary, over last year one could observe a continuous hardening of the US' China policy - including granting more political and military support to the Taiwan island, encouraging US allies and partners in Asia to increase their defense spending, engaging in more navel activities in the Pacific and imposing more technology sanctions on China. In the meantime, economic and social problems within the US are mounting. The national debt ceiling is only the tip of an iceberg - the future of the American economy is now clouded by high US Federal Reserve interest rates that slow down growth, feed unemployment and might well lead to a recession. Moreover, the US society remains split along the same lines it was during the presidency of Donald Trump. The Biden administration has clearly failed to reunite America: Many of the social, political, regional, ethnic and even generational divisions have got only deeper since January 2021. It is hard to imagine how a nation divided so deeply and along so many lines could demonstrate continuity and strategic vision in its foreign policy, or to allocate financial resources needed to sustain a visionary and consistent global leadership. Of course, the "dual containment" policy is not the only illustration of the gap between the US ambitions and its resources. The same gap inevitably pops up at every major forum that the US conducts with select groups of countries from the Global South - Africa, Southeast Asia, Latin America or the Middle East. The Biden administration has no shortage of arguments warning these countries about potential perils of cooperating with Moscow or Beijing, but it does not offer too many plausible alternatives that would showcase the US generosity, its strategic vision, and its true commitment to the burning needs of the US interlocutors. To cut it short, Uncle Sam brings lots of sticks to such meetings, but not enough carrots to win the audience. In sum, US foreign policy under President Joe Biden reminds people of a very advanced and highly sophisticated smartphone that has a rather weak battery, which is not really energy efficient. The proud owner of the gadget has to look perennially for a power socket in order not to have the phone running out of power at any inappropriate moment. Maybe the time has come for the smartphone owner to look for another model that would have fewer fancy apps, but a stronger and a more efficient battery, which will make the appliance more convenient and reliable.