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
Portadown businessman avoids jail for sexual assault of teen under his employment
Defence said the defendant 'continues to deny' the charges and bail in the sum of £1,000 was fixed for appeal
A Portadown man has avoided jail after sexually assaulting a 16-year-old shop worker under his employment. -ADVERTISEMENT- Brian Thomas Chapman (58), of Moyallan Road, appeared before Newry Magistrates’ Court on Monday for sentencing on two counts of sexual assault. The prosecution outlined that on September 23, 2020, a 16-year-old student in the employment of Brian Chapman, disclosed to her mother about incidents that had occurred in her workplace. She said Chapman had put his hand on her thigh and the back of her leg. She also disclosed that she had been getting extra money from him and he had been sending her text messages. The allegations were reported to police the next day, September 24. The victim then took part in an interview on October 9, in which she said, when she was alone in Chapman’s office, he placed his hand on her upper thigh and his other hand on her lower back, underneath her trousers. The defendant was arrested and interviewed at Lurgan police station, where he denied the allegations. His phone was seized and an examination was carried out. The first interview of the defendant took place on October 9, during which he admitted to sending a message about wanting the victim to work 24/7, but stated this was a joke. The second interview took place on January 28, 2021, where he admitted to sending the 24/7 message, but denied sending other messages, such as “hope you’re spending the pounds on something special”. Throughout this process, Chapman denied sending the messages and denied any of the sexual assaults alleged by the victim. On the Chapman’s criminal record, the prosecution added that he was convicted of three common assaults on appeal. In terms of commission, these matters pre-dated this case but the conviction occurred during the running of this case and also involved a female working for the defendant. Prosecution continued that the age of the victim was an aggravating feature, arguing there was a “vulnerability” due to the “power-imbalance” between Chapman and the young student working for him. An additional aggravating feature, they said, was that during the course of the defence, part of the defence was that the victim had “manipulated or manufactured” some of the text messages that were sent. A defence lawyer, speaking on the pre-sentence report, noted the author deemed Chapman to be of low risk. He also noted that similar offences were contested in the County Court in respect of another complaint, with the judge substituting indecent assault charges for common assault. He also argued a Sexual Offences Prevention Order (SOPO) was not necessary as the offending was four years ago, there has been no repetition and risk had been addressed. District Judge Eamonn King noted the defendant was convicted on two of four original charges following a contest, which ran over a number of days, with the case adjourned for a pre-sentence report and victim impact statement to be produced. He added the defendant “continues to deny” the charges and seeks to appeal the outcome. District Judge King, on reading the pre-sentence report, noted the defendant “denies ever hugging or touching the individual and he denies any sexual attraction to the victim”, but pointed to a paragraph in the report which stated, “From the available evidence, it’s possible to surmise that he demonstrated risk taking and impulsive behaviour. It appears that he took advantage of his position and power in a bid to meet his sexual needs, given the victim’s young age and the fact that he was her employer”. The report added that this demonstrated “limited victim empathy and responsibility due to his denial of the offences”. On the victim impact statement, District Judge King described her as a young girl getting her first job, with the “world as her oyster”. He continued: “As a result of what she says occurred, that turned on its head. It left her feeling inwardly uncomfortable, anxious and lonely. She cut herself off from her friends. She stopped going out. She didn’t want to go to school.” He also described a “degree of manipulation” in the case, as this was the victim’s first job and there was a power imbalance between her as an employee, and Chapman as the employer. In his sentencing remarks, District Judge King, said: “I’ve taken time to emphasise to the victim in this case that the victim did nothing wrong. The victim did everything right and the victim shouldn’t feel lonely, anxious or isolated. “The victim should feel confident, strong and outgoing.” Owing to the defendant’s ongoing denial of the charges, he added: “My sentencing exercise isn’t the conclusion of the case today, but I will sentence, so that we can move towards the conclusion going forward. “I am satisfied, irrespective of what the pre-sentence report says, that the defendant took advantage of someone, attempted to groom someone and was guilty of the two offences.” On the two counts, Chapman was sentenced to three months in prison, suspended for two years. He was also made subject to a Sexual Offences Prevention Order (SOPO) for five years and placed on the sex offenders’ register for seven years. Following sentencing, District Judge King fixed bail for appeal at £1,000.
United Airlines Boeing 757 loses tire during takeoff
United Airlines confirmed that a tire fell off a Boeing 757 passenger plane when it took off from Los Angeles International Airport in the early morning of July 8, local time. It is reported that there has been no report of any material damage or casualties in this accident. United Airlines said in a statement that the aircraft tire has been found and the investigation is ongoing. There were 174 passengers and 7 crew members on the flight involved. The flight left Los Angeles International Airport at around 7:15 on July 8 and flew to Denver. Unlike the aircraft tire falling incident on a United Airlines flight in March, the pilot of this flight continued the journey to Denver and landed smoothly there.
Ukrainian Presidential Office: Russia's attacks on multiple locations in Ukraine have killed 36 people
The Ukrainian presidential office said on July 8 local time that Russia's large-scale attacks on many parts of Ukraine have killed 36 people and injured 140 others. According to the Ukrainian State Emergency Service, a total of 619 rescue workers and 132 equipment participated in the rescue work across Ukraine that day. Ukrainian President Zelensky said on social media on the 8th that Russia launched more than 40 missiles of various types at Ukraine that day. Residential buildings and infrastructure in many cities in Ukraine were damaged in varying degrees of attacks, and a children's hospital was destroyed. Rescue departments are currently conducting emergency rescue on the scene. The Russian Ministry of Defense issued a statement on the 8th local time saying that Ukrainian officials' claim that Russia used missiles to attack Ukrainian civilian facilities was untrue. The damage suffered by Kiev was caused by the fall of missiles launched by the city's air defense system.
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.
Google may bring Google Wallet for Indian users
Google Wallet can help you store your IDs, driving license, loyalty cards, concert tickets and more. You can also store your payment cards and use tap to pay to pay anywhere Google Pay is accepted. Google wallet is available in various countries but Google never launched it in India. Google let indian users stick with the Gpay which facilitates UPI payments. Tap to pay is not part of it. Also we can not store things such as IDs and Passes in indian version of Gpay. This might change and Google may launch Google Wallet in India. With the recent version of Google Wallet and Google Play Services, Google has added some flags and code which indicate that Google is working on something for Indian users regarding wallet. The first change I noticed recently when going through the Google Play Services apk was addition of two new flags Both flags are part of com.google.android.gms.pay package in the Google Play Services. This package contains all the flags for features of Gpay/Wallet. Google does server side flipping of flags to enable/disable features for users. So both these flags doesn't really provide any info about what features enabling these flags is going to bring. But the point here is that Google Wallet is not launched in India so why Google added these flags inside Play Services ? The answer could be that Google may be working on bringing Google Wallet to India. It can enable tap to pay, store payments and various other features for Indian users which we don't have in the current Gpay for India. I found similar flags in the analysis Google Wallet APK - These flags are also disabled by default. But this is again a clear indication of Google working towards something for Indian users. In both cases, enabling the flags doesn't bring anything noticeable UI or feature because there is nothing much added besides flags. Google has dogfood/testing versions internally, so the code will show up slowly in upcoming versions. The last piece of code I found is also from Google Play Services. In case you don't know, Google was working on Digilocker integration in the Google Files app which was supposed to bring your digital document inside the app such as driving license, COVID certificates, aadhar card. But Google has ditched the effort of bringing these features and they removed the "Important" tab (where digilocker was supposed to be integrated) from the Google Files app completely. So things are going to change and here is how. This is the code which I found in the Google Play Services - So the word "PASS" along with PAN, DRIVERS LICENCE, VACC CERTIFICATE & AADHAR CARD, is clear indication of the possibility of Google adding support for these directly through Google Wallet using Digilocker, just like Samsung Pass does it. This code is not old as I have checked older beta versions of Play Services where this code is not present. Here is a string which was added in a previous beta version a few weeks ago but I completely ignored it because it didn't make any sense without flags and the other code - This addition was surprising because there was nothing regarding digilocker before in the Play Services. In the words "pay_valuable", the "pay" to Wallet/Gpay and "valuable" refers to the things like Passes, loyalty cards and transit cards. Since we are talking about digilocker, these "valuable" are driving license, vaccination certificate, PAN card and Aadhar card which can be store in Google Wallet after digilocker integration. That's all about it. We will know more about it in upcoming app updates or maybe Google can itself annouce something about this.