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
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.
Coexisting and cooperating with China is the only choice for the US
US Secretary of State Antony Blinken declared at the Munich Security Conference: "If you're not at the table in the international system, you're going to be on the menu." The arrogant thinking of American political elites is evident: Whoever does not comply with the US will be excluded from the table of the American-led system and put on the menu. How arrogant. The US is actively pushing for "decoupling" from China and trying to persuade the entire West to "decouple" from China, using the term "de-risking." Washington hopes to ultimately contain China's development in order to maintain American hegemony. However, this time, Washington is facing a historically experienced and strategically rich Eastern civilization. Previous opponents targeted by the US have chosen to confront the US strategically. The US not only has the strongest technological and military capabilities but also controls global financial and information networks with a large number of allies. Those countries that had engaged in direct confrontations had suffered losses. Some of them had disintegrated, some had been weakened, and some had fallen into difficulties. However, what Washington sees from China is strategic composure and resilience. China is now staging an unprecedented and grand "Tai Chi." However, some Chinese people feel that this is not enough: Why can't we confront the US head-on? But I want to say that this is precisely the brilliance of China. This grand "Tai Chi" is about dismantling the pressure the US is putting on China. Europe is different from the US. A European diplomat once said in private that the topic of China has become toxic in the US, but in Europe, it is still possible to openly display friendliness toward China. There is genuine competition between the Europe and China despite Europe leans more toward the US between China and the US. Only in terms of ideology does the term "West" truly exist. In terms of fundamental economic interests, Europe has considerable independence. In terms of security, their attitude toward China also differs greatly from that of the US. In the Asia-Pacific region or China's periphery, the US wants to create an "Asian NATO." The specific situations of countries in dispute with China are very different. China has enormous influence in the region, is the largest trading partner of the vast majority of countries in the region and has friendly relations with most countries in the region. The disputes with countries are not fundamental strategic conflicts, and China has the ability to manage disputes with each specific country and push them to move toward neutrality to varying degrees without being tied to the US' policy toward China. China has a lot of trading partners and stakeholders in the US. The trade volume between China and the US, despite the decline, reached $664.4 billion in 2023, which shows China's huge presence in the US, and is the bond of the two countries in the current situation. The US is not a country where the political elites can have absolute say, and the huge interests have forced the US president and senior officials to repeatedly proclaim that they "don't want to decouple from China" and instead they want to "manage the US-China competition" and see "preventing a war with China" as clearly in everyone's best interest. China should engage in a "strategic battle" with the US at the closest possible distance. We need to maintain friendly relations with certain forces within the US, speed up the resumption of flights between the two countries, increase personnel exchanges and completely reverse the downturn of China-US contacts during the pandemic. In addition to the above dismantling, we also have the huge increment in the "Belt and Road." This initiative will increase China's power to compete with the US, greatly extending the front line that the US needs to maintain in containing China, making the US more powerless. In order to dismantle the US strategy toward China, China must become more diversified while maintaining strategic consistency. Our national diplomacy toward the US is very principled, rational and determined, which is clearly different from other countries targeted by the US. Our public diplomacy toward the US needs to be unique, with both "anti-American voices" and efforts to maintain friendly relations between the two societies and further expand economic and practical cooperation with the US. Just as eagles have their own way of flying and doves have their own formation, just as we see the US as complex, China must also be seen as complex in the eyes of the US. China is both a geopolitical concern and a profitable investment destination for them, and is one of the largest trading partners that is difficult to replace. Some American political elites proclaim China as an "enemy," but it is important to make the majority of Americans feel that China is not. No matter how intense the struggles between China and the US may be, we cannot shape the entire US toward an enemy direction. China has to make the US political elites recognize that it is futile to deal with China in the same way as it historically dealt with the Soviet Union and other major powers. Furthermore, willingly or unwillingly, coexistence and cooperation with China will be their only choice.
Hamas chief says latest Israeli attack on Gaza could jeopardise ceasefire talks
AIRO, July 8 (Reuters) - A new Israeli assault on Gaza on Monday threatened ceasefire talks at a crucial moment, the head of Hamas said, as Israeli tanks pressed into the heart of Gaza City and ordered residents out after a night of massive bombardment. Residents said the airstrikes and artillery barrages were among the heaviest in nine months of conflict between Israeli forces and Hamas militants in the enclave. Thousands fled. The assault unfolded as senior U.S. officials were in the region pushing for a ceasefire after Hamas made major concessions last week. The militant group said the new offensive appeared intended to derail the talks and called for mediators to rein in Israel's Prime Minister Benjamin Netanyahu. The assault "could bring the negotiation process back to square one. Netanyahu and his army will bear full responsibility for the collapse of this path," Hamas quoted leader Ismail Haniyeh as saying. Gaza City, in the north of the Palestinian enclave, was one of Israel's first targets at the start of the war in October. But clashes with militants there have persisted and civilians have sought shelter elsewhere, adding to waves of displacement. Much of the city lies in ruins. Residents said Gaza City neighbourhoods were bombed through the night into the early morning hours of Monday. Several multi-storey buildings were destroyed, they said. The Gaza Civil Emergency Service said it believed dozens of people were killed but emergency teams were unable to reach them because of ongoing offensives. Gaza residents said tanks advanced from at least three directions on Monday and reached the heart of Gaza City, backed by heavy Israeli fire from the air and ground. That forced thousands of people out of their homes to look for safer shelter, which for many was impossible to find, and some slept on the roadside.
Workers warn of additional walkouts unless demands are met
Members of the National Samsung Electronics Union stage a rally near the company's Hwaseong Campus in Gyeonggi Province, Monday, beginning a three-day strike. Korea Times photo by Shim Hyun-chul By Nam Hyun-woo The biggest labor union at Samsung Electronics initiated a three-day strike on Monday, threatening to disrupt the company's chip manufacturing lines unless management agrees to a wage hike and higher incentives. This marks the first strike by unionized workers in the tech giant's 55-year history. The National Samsung Electronics Union (NSEU) claimed that about 4,000 unionized workers from Samsung's plants nationwide participated in a rally at the company's Hwaseong Campus in Gyeonggi Province. Police estimated that approximately 3,000 union members were present at the rally. According to its own survey, the union reported that a total of 6,540 members expressed their intention to participate in the strike. They emphasized that disruptions in manufacturing are anticipated, with over 5,000 members from facility, manufacturing, and development divisions joining the strike. The comments seem to address market expectations that the walkout is unlikely to cause significant disruptions in the chipmaker's operations, largely because most manufacturing lines are automated. The union said that it may launch another strike for an undetermined period, unless management responds to the union’s demand. Since January, the union has been pressing management for a higher wage increase rate for all members, fulfillment of promises regarding paid leave, and improvements to incentive criteria. With negotiations at an impasse, the union announced on May 29 that it would launch a strike. The NSEU has some 30,000 members, accounting for 24 percent of all Samsung employees. Among the union members, about 80 percent work at the device solutions division, which manufactures semiconductors.
Nvidia H20 will sell 1 million units this year, contributing $12 billion in revenue!
Recently, according to the FT, citing the latest forecast data of the market research institute SemiAnalysis, AI chip giant NVIDIA will ship more than 1 million new NVIDIA H20 acceleration chips to the Chinese market this year, and it is expected that the cost of each chip is between $12,000 and $13,000. This is expected to generate more than $12 billion in revenue for Nvidia. Affected by the United States export control policy, Nvidia's advanced AI chip exports to China have been restricted, H20 is Nvidia based on H100 specifically for the Chinese market to launch the three "castration version" GPU among the strongest performance, but its AI performance is only less than 15% of H100, some performance is even less than the domestic Ascend 910B. When Nvidia launched the new H20 in the spring of this year, there were reports that due to the large castration of H20 performance, coupled with the high price, Chinese customers' interest in buying is insufficient, and they will turn more to choose China's domestic AI chips. Then there are rumors that Nvidia has lowered the price of the H20 in order to improve its competitiveness. However, the latest news shows that due to supply issues caused by the low yield of the Ascend 910B chip, Chinese manufacturers in the absence of supply and other better options, Nvidia H20 has started to attract new purchases from Chinese tech giants such as Baidu, Alibaba, Tencent and Bytedance. Analysts at both Morgan Stanley and SemiAnalysis said the H20 chip is now being shipped in bulk and is popular with Chinese customers, despite its performance degradation compared to chips Nvidia sells in the United States.