Technology

MM1, a Family of Multimodal AI Models with up to 30 billion Parameters, is being Developed by Apple Researchers

Published

8 months ago

March 18, 2024

In a pre-print paper, Apple researchers presented their work on developing a multimodal large language model (LLM) for artificial intelligence (AI). The paper describes how it was possible to achieve the advanced capabilities of multimodality and train the foundation model on both text-only data and images, and it was published on an online portal on March 14. The Cupertino-based tech giant has made new advances in AI in response to CEO Tim Cook’s statement during the company’s earnings calls, which stated that AI features might be released later this year.

ArXiv, an open-access online repository for scholarly papers, has published the research paper’s pre-print version. Peer review is not, however, applied to the papers that are posted here. The project is thought to be connected to Apple as well, even though the paper makes no mention of the company; this is because the majority of the researchers mentioned are connected to the machine learning (ML) division of Apple.

A family of multimodal models with up to 30 billion parameters, known as MM1, is the project that the researchers are currently working on. The paper’s authors referred to it as a “performant multimodal LLM (MLLM)” and noted that in order to build an AI model that can comprehend both text and image-based inputs, image encoders, the vision language connector, and other architecture elements and data decisions were made.

The paper provided an example in stating that “We demonstrate that achieving state-of-the-art (SOTA) few-shot results across multiple benchmarks, compared to other published pre-training results, requires a careful mix of image-caption, interleaved image-text, and text-only data for large-scale multimodal pre-training.”

To put it simply, the AI model has not received enough training to produce the intended results and is presently in the pre-training phase. This phase involves designing the model’s workflow and data processing eventually using the algorithm and AI architecture. The researchers at Apple were able to incorporate computer vision into the model by means of a vision language connector and image encoders. Upon conducting tests using a combination of image-only, image-text, and text-only data sets, the team discovered that the outcomes were comparable to those of other models at the same stage.

Although this is a significant breakthrough, there is insufficient evidence in this research paper to conclude that Apple will integrate a multimodal AI chatbot into its operating system. It’s difficult to even say at this point whether the AI model is multimodal in terms of receiving inputs or producing output (i.e., whether it can produce AI images or not). However, it can be said that the tech giant has made significant progress toward developing a native generative AI foundation model if the results are verified to be consistent following peer review.

Up Next

NVIDIA Releases 6G Research Cloud Platform to Use AI to Improve Wireless Communications

Don't Miss

Google Gemini may be used by Apple to Power Certain AI Features on the iPhone

Kajal Chavan

Technology

Google is said to be discontinuing the Pixel Tablet 2 and may be leaving the market once more

Published

45 minutes ago

November 21, 2024

Archana Suryawanshi

Google terminated the development of the Pixel Tablet 3 yesterday, according to Android Headlines, even before a second-generation model was announced. The second-generation Pixel Tablet has actually been canceled, according to the report. This means that the gadget that was released last year will likely be a one-off, and Google is abandoning the tablet market for the second time in just over five years.

If accurate, the report indicates that Google has determined that it is not worth investing more money in a follow-up because of the dismal sales of the Pixel Tablet. Rumors of a keyboard accessory and more functionality for the now-defunct project surfaced as recently as last week.

It’s important to keep in mind that Google’s Nest subsidiary may abandon its plans for large-screen products in favor of developing technologies like the Nest Hub and Hub Max rather than standalone tablets.

Google has always had difficulty making a significant impact in the tablet market and creating a competitor that can match Apple’s iPad in terms of sales and general performance, not helped in the least by its inconsistent approach. Even though the hardware was good, it never really fought back after getting off to a promising start with the Nexus 7 eons ago. Another problem that has hampered Google’s efforts is that Android significantly trails iPadOS in terms of the quantity of third-party apps that are tablet-optimized.

After the Pixel Slate received tremendously unfavorable reviews, the firm first declared that it was finished producing tablets in 2019. Two tablets that were still in development at the time were discarded.

By 2022, however, Google had altered its mind and declared that a tablet was being developed by its Pixel hardware team. The $499 Pixel Tablet was the final version of the gadget, which came with a speaker dock that the tablet could magnetically connect to. (Google would subsequently charge $399 for the tablet alone.)

Technology

Windows 11 PCs with Arm Processors now have an Official ISO for Clean Installations

Published

7 days ago

November 15, 2024

Archana Suryawanshi

Power users occasionally prefer to start over when they acquire a new computer, so they follow the pro-gamers’ advice and reinstall Windows using a brand-new ISO image that comes straight from Microsoft and is free of bloatware and needlessly complex “driver management programs.” Up until recently, the new Snapdragon laptops’ more specialized version of Windows 11 didn’t support that.

The Windows 11 build on these new laptops is unusual because of the Arm64-based hardware, which differs from the typical x86 and x64 innards found in most laptops and desktops. Microsoft has finally released a disk image (or ISO file) for these devices after several months of waiting. To perform a direct reinstallation or make a bootable flash drive for a different device, you may now download it straight from Microsoft’s website. It is identical to the installation media utility that is currently available.

Be aware that there may be some glitches if you use this method for a fresh install. Compared to previous designs, the Snapdragon X system-on-a-chip has a lot fewer hardware variables, but because it’s so new, Windows Update might not include all the necessary components. You may need to use an Ethernet connection or the old-fashioned sneakernet to manually load drivers from another computer. You may also need to do some Googling to locate all the files you require for that.

Technology

OPPO Reno 13 series will debut in China shortly, with India following in 2025

Published

1 week ago

November 14, 2024

Archana Suryawanshi

According to reports, OPPO, a Chinese firm, is getting ready to introduce its Reno 13 series smartphones in its native nation this month. As per 91Mobiles, the OPPO Reno 13 and Reno 13 Pro models are anticipated to debut in China on November 25. The Indian launch is probably set for January 2025. The smartphone series that debuted in July of this year, the Reno 12 series, will be replaced by the Reno 13 series.

Information regarding the specifications of the new Reno 13 and Reno 13 Pro smartphones has leaked online, although the business has not yet confirmed the launch date. These are the specifics:

OPPO Reno 13 Series: Anticipations

It is anticipated that the OPPO Reno 13 Pro would have a 6.78-inch, quad-curved OLED screen with 1.5K resolution. In contrast, the slightly smaller 6.7-inch display with FHD+ resolution is found on the OPPO Reno 12 Pro. In China, the Pro model is probably going to be powered by the MediaTek Dimensity 8350 chipset, while in India, it might have a different processor. A 50MP primary camera, an 8MP ultrawide sensor, and a 50MP telephoto sensor with 3x optical zoom are anticipated to be included in the OPPO Reno 13 Pro’s photographic setup. Most likely, the front camera will include a 50MP sensor.

With a 5,900mAh battery as opposed to the 5,000mAh battery on the Reno 12 Pro, the Reno 13 Pro is anticipated to significantly increase battery capacity. Additionally, it is anticipated that the smartphone would support both 50W wireless and 80W wired charging. Additionally, an IP68/IP69 designation for water and dust protection could increase its durability.

Although the price of the smartphones in the Reno 13 series is not well known, it is anticipated to be similar to that of its predecessor. For comparison, the 12GB RAM + 256GB storage version of the OPPO Reno 12 Pro launched at Rs 36,999, while the 8GB RAM + 256GB storage version of the vanilla model cost Rs 32,999.

OPPO Reno 13 Pro: Anticipated features

Display: 6.78-inch OLED, quad-curved, with a refresh rate of 120 Hz and a resolution of 1.5K
processor: MediaTek Dimensity 8350
rear camera: 50MP primary, 8MP ultra-wide, and 50MP telephoto (3x zoom)
front camera: 50MP
Battery: 5,900mAh
Charging: 50W wireless and 80W wired
IP rating: IP68/IP69; operating system: ColorOS 15 based on Android 15