The A.I Megathread (LLM , GPT , Development)

[bnew]

Animate Anyone

Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation

humanaigc.github.io

Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation​

Li Hu, Xin Gao, Peng Zhang, Ke Sun, Bang Zhang, Liefeng Bo

Institute for Intelligent Computing，Alibaba Group

Paper video Code arXiv​

https://humanaigc.github.io/animate-anyone/static/videos/teaser1.mp4

Abstract​

Character Animation aims to generating character videos from still images through driving signals. Currently, diffusion models have become the mainstream in visual generation research, owing to their robust generative capabilities. However, challenges persist in the realm of image-to-video, especially in character animation, where temporally maintaining consistency with detailed information from character remains a formidable problem. In this paper, we leverage the power of diffusion models and propose a novel framework tailored for character animation. To preserve consistency of intricate appearance features from reference image, we design ReferenceNet to merge detail features via spatial attention. To ensure controllability and continuity, we introduce an efficient pose guider to direct character's movements and employ an effective temporal modeling approach to ensure smooth inter-frame transitions between video frames. By expanding the training data, our approach can animate arbitrary characters, yielding superior results in character animation compared to other image-to-video methods. Furthermore, we evaluate our method on benchmarks for fashion video and human dance synthesis, achieving state-of-the-art results.​

Method​

The overview of our method. The pose sequence is initially encoded using Pose Guider and fused with multi-frame noise, followed by the Denoising UNet conducting the denoising process for video generation. The computational block of the Denoising UNet consists of Spatial-Attention, Cross-Attention, and Temporal-Attention, as illustrated in the dashed box on the right. The integration of reference image involves two aspects. Firstly, detailed features are extracted through ReferenceNet and utilized for Spatial-Attention. Secondly, semantic features are extracted through the CLIP image encoder for Cross-Attention. Temporal-Attention operates in the temporal dimension. Finally, the VAE decoder decodes the result into a video clip.

---

GitHub - HumanAIGC/AnimateAnyone: Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation

Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation - HumanAIGC/AnimateAnyone

github.com

AnimateAnyone​

Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation

Li Hu, Xin Gao, Peng Zhang, Ke Sun, Bang Zhang, Liefeng Bo

https://arxiv.org/pdf/2311.17117.pdf

 

[bnew]

Google DeepMind's AI Dreamed Up 380,000 New Materials. The Next Challenge Is Making Them

Google DeepMind researchers say they’ve expanded the number of known stable materials tenfold. Some could be useful for everything from batteries to superconductors—if they make it out of the lab.

www.wired.com

https://archive.is/GL4vJ

Google DeepMind's AI Dreamed Up 380,000 New Materials. The Next Challenge Is Making Them​

Google DeepMind researchers say they’ve expanded the number of known stable materials tenfold. Some could be useful for everything from batteries to superconductors—if they make it out of the lab.

https://media.wired.com/clips/6566927a0f92a312029410f1/720p/pass/a-lab%20b-roll-1.mp4

A-Lab in February of 2023 at Lawrence Berkeley National Laboratory in Berkeley, California.
VIDEO: MARILYN SARGENT/BERKELEY LAB

The robotic line cooks were deep in their recipe, toiling away in a room tightly packed with equipment. In one corner, an articulated arm selected and mixed ingredients, while another slid back and forth on a fixed track, working the ovens. A third was on plating duty, carefully shaking the contents of a crucible onto a dish. Gerbrand Ceder, a materials scientist at Lawrence Berkeley National Lab and UC Berkeley, nodded approvingly as a robotic arm delicately pinched and capped an empty plastic vial—an especially tricky task, and one of his favorites to observe. “These guys can work all night,” Ceder said, giving two of his grad students a wry look.

Stocked with ingredients like nickel oxide and lithium carbonate, the facility, called the A-Lab, is designed to make new and interesting materials, especially ones that might be useful for future battery designs. The results can be unpredictable. Even a human scientist usually gets a new recipe wrong the first time. So sometimes the robots produce a beautiful powder. Other times it’s a melted gluey mess, or it all evaporates and there’s nothing left. “At that point, the humans would have to make a decision: What do I do now?” Ceder says.

The robots are meant to do the same. They analyze what they’ve made, adjust the recipe, and try again. And again. And again. “You give them some recipes in the morning and when you come back home you might have a nice new soufflé,” says materials scientist Kristin Persson, Ceder’s close collaborator at LBNL (and also spouse). Or you might just return to a burned-up mess. “But at least tomorrow they’ll make a much better soufflé.”

https://media.wired.com/clips/65669276f082e4efb66daf9b/720p/pass/a-lab-b-roll-2.mp4

PLAY/PAUSE BUTTONVIDEO: MARILYN SARGENT/BERKELEY LAB

Recently, the range of dishes available to Ceder’s robots has grown exponentially, thanks to an AI program developed by Google DeepMind. Called GNoME, the software was trained using data from the Materials Project, a free-to-use database of 150,000 known materials overseen by Persson. Using that information, the AI system came up with designs for 2.2 million new crystals, of which 380,000 were predicted to be stable—not likely to decompose or explode, and thus the most plausible candidates for synthesis in a lab—expanding the range of known stable materials nearly 10-fold. In a paper published today in Nature, the authors write that the next solid-state electrolyte, or solar cell materials, or high-temperature superconductor, could hide within this expanded database.

Finding those needles in the haystack starts off with actually making them, which is all the more reason to work quickly and through the night. In a recent set of experiments at LBNL, also published today in Nature, Ceder’s autonomous lab was able to create 41 of the theorized materials over 17 days, helping to validate both the AI model and the lab’s robotic techniques.

When deciding if a material can actually be made, whether by human hands or robot arms, among the first questions to ask is whether it is stable. Generally, that means that its collection of atoms are arranged into the lowest possible energy state. Otherwise, the crystal will want to become something else. For thousands of years, people have steadily added to the roster of stable materials, initially by observing those found in nature or discovering them through basic chemical intuition or accidents. More recently, candidates have been designed with computers.

The problem, according to Persson, is bias: Over time, that collective knowledge has come to favor certain familiar structures and elements. Materials scientists call this the “Edison effect,” referring to his rapid trial-and-error quest to deliver a lightbulb filament, testing thousands of types of carbon before arriving at a variety derived from bamboo. It took another decade for a Hungarian group to come up with tungsten. “He was limited by his knowledge,” Persson says. “He was biased, he was convinced.”

FEATURED VIDEO
Harvard Professor Explains Algorithms in 5 Levels of Difficulty

 

[bnew]

{continued}

DeepMind’s approach is meant to look beyond those biases. The team started with 69,000 materials from Persson’s library, which is free to use and funded by the US Department of Energy. That was a good start, because the database contains the detailed energetic information needed to understand why some materials are stable and others aren’t. But it wasn’t enough data to overcome what Google DeepMind researcher Ekin Dogus Cubuk calls a “philosophical contradiction” between machine learning and empirical science. Like Edison, AI struggles to generate truly novel ideas beyond what it has seen before. “In physics, you never want to learn a thing that you already know,” he says. “You almost always want to generalize out of domain”—whether that’s to discover a different class of battery material or a new superconductivity theory.

GNoME relies on an approach called active learning. First, an AI called a graph neural network, or GNN, uses the database to learn patterns in the stable structures and figure out how to minimize the energy in the atomic bonds within new structures. Using the whole range of the periodic table, it then produces thousands of potentially stable candidates. The next step is to verify and adjust them, using a quantum mechanics technique called density-functional theory, or DFT. These refined results are then plugged back into the training data and the process is repeated.

The structures of 12 compounds in the Materials Project database.
ILLUSTRATION: JENNY NUSS/BERKELEY LAB


The researchers found that, with multiple repetitions, this approach could generate more complex structures than were initially in the Materials Project data set, including some that were composed of five or six unique elements. (The data set used to train the AI largely capped out at four.) Those types of materials involve so many complex atomic interactions that they generally escape human intuition. “They were hard to find,” Cubuk says. “But now they’re not so hard to find anymore.”

But DFT is only a theoretical validation. The next step is actually making something. So Ceder’s team picked 58 crystals to create in the A-Lab. After taking into account the capabilities of the lab and available precursors, it was a random selection. And at first, as expected, the robots failed, then repeatedly adjusted their recipes. After 17 days of experiments, the A-Lab managed to produce 41 of the materials, or 71 percent, sometimes after trying a dozen different recipes.

Taylor Sparks, a materials scientist at the University of Utah who wasn’t involved in the research, says that it’s promising to see automation at work for new types of materials synthesis. But using AI to propose thousands of new hypothetical materials, and then chasing after them with automation, just isn’t practical, he adds. GNNs are becoming widely used to develop new ideas for materials, but usually researchers want to tailor their efforts to produce materials with useful properties—not blindly produce hundreds of thousands of them. “We’ve already had way too many things that we’ve wanted to investigate than we physically could,” he says. “I think the challenge is, is this scaled synthesis approaching the scale of the predictions? Not even close.”

Only a fraction of the 380,000 materials in the Nature paper will likely wind up being practical to create. Some involve radioactive elements, or ones that are too expensive or rare. Some will require types of synthesis that involve extreme conditions that can’t be produced in a lab, or precursors that lab suppliers don’t have on hand.

That’s likely even true for materials that could very well hold potential for the next photovoltaic cell or battery design. “We've come up with a lot of cool materials,” Persson says. “Making them and testing them has consistently been the bottleneck, especially if it's a material that nobody's ever made before. The number of people I can call up in my circle of friends who go, ‘Absolutely, let me get on that for you,’ is pretty much one or two people.’”

“Really, is it that high?” Ceder interjects with a laugh.

Even if a material can be made, there’s a long road to turning a basic crystal into a product. Persson brings up the example of an electrolyte inside a lithium-ion battery. Predictions about the energy and structure of a crystal can be applied to problems like figuring out how easily lithium ions can move across it—a key aspect of performance. What it can’t predict as easily is whether that electrolyte will react with neighboring materials and destroy the whole device. Plus, in general, the utility of new materials only becomes apparent in combination with other materials or by manipulating them with additives.

Still, the expanded range of materials expands the possibilities for synthesis, and also provides more data for future AI programs, says Anatole von Lilienfeld, a materials scientist at the University of Toronto who wasn’t involved in the research. It also helps nudge materials scientists away from their biases and towards the unknown. “Every new step that you take is fantastic,” he says. “It could usher in a new compound class.”

https://media.wired.com/clips/65668a6f2838afae9c95573a/720p/pass/Video_Google_Materials_9seconds%20copy.mp4

PLAY/PAUSE BUTTON
The Materials Project can visualize the atomic structure of materials. This compound (Ba₆Nb₇O₂₁) is one of the new materials calculated by GNoME. It contains barium (blue), niobium (white), and oxygen (green).VIDEO: MATERIALS PROJECT/BERKELEY LAB


Google is also interested in exploring the possibilities of the new materials generated by GNoME, says Pushmeet Kohli, vice president of research at Google DeepMind. He compares GNoME to AlphaFold, the company’s software that startled structural biologists with its success at predicting how proteins fold. Both are addressing fundamental problems by creating an archive of new data that scientists can explore and expand. From here, the company plans to work on more specific problems, he says, such as homing in on interesting material properties and using AI to speed up synthesis. Both are challenging problems, because there is typically far less data to start with than there is for predicting stability.

Kohli says the company is exploring its options for working more directly with physical materials, whether by contracting outside labs or continuing with academic partnerships. It could also set up its own lab, he adds, referring to Isomorphic Labs, a drug discovery spinoff from DeepMind established in 2021 following the success of AlphaFold.

Things could get complicated for researchers trying to put the materials to practical use. The Materials Project is popular with both academic labs and corporations because it allows any type of use, including commercial ventures. Google DeepMind’s materials are being released under a separate license that forbids commercial use. “It’s released for academic purposes,” Kohli says. “If people want to investigate and explore commercial partnerships, and so on, we will review them on a case-by-case basis.”

Multiple scientists who work with new materials noted that it’s unclear what sort of say the company would have if testing in an academic lab led to a possible commercial use for a GNoME-generated material. An idea for a new crystal—without a particular use in mind—is generally not patentable, and tracing its provenance back to the database could be difficult.

Kohli also says that while the data is being released, there are no current plans to release the GNoME model. He cites safety considerations—the software could theoretically be used to dream up dangerous materials, he says—and uncertainty about Google DeepMind’s materials strategy. “It is difficult to make predictions about what the commercial impact would be,” Kohli says.

Sparks expects his fellow academics to bristle at the lack of code for GNoME, just as biologists did when AlphaFold was initially published without a complete model. (The company later released it.) “That’s lame,” he says. Other materials scientists will likely want to reproduce the results and investigate ways to improve the model or tailor it to specific uses. But without the model, they can’t do either, Sparks says.

In the meantime, the Google DeepMind researchers hope hundreds of thousands of new materials will be enough to keep theorists and synthesizers—both human and robotic—plenty busy. “Every technology could be improved with better materials. It’s a bottleneck,” Cubuk says. “This is why we have to enable the field by discovering more materials, and helping people discover even more.”

 

[bnew]

Spoiler: more tweets

 

[bnew]

https://archive.is/mGrNQ

 

[bnew]

https://archive.is/m6yyC

Spoiler: more tweets

 

[YvrzTrvly]

So essentially...considering this thread is an abortion in regards to formatting...we're doing nothing beneficial with this tech besides usurpation and visual masturbation. Got it. Wake me up when John Connor needs to be saved

 

[bnew]

GitHub - lucidrains/q-transformer: Implementation of Q-Transformer, Scalable Offline Reinforcement Learning via Autoregressive Q-Functions, out of Google Deepmind

Implementation of Q-Transformer, Scalable Offline Reinforcement Learning via Autoregressive Q-Functions, out of Google Deepmind - lucidrains/q-transformer

github.com

 

[bnew]

 

[bnew]

 

[bnew]

How Googlers cracked OpenAI's ChatGPT with a single word

A research team from Bay Area tech giant Google got OpenAI's ChatGPT to spit out its private training data in a new study.

www.sfgate.com

How Googlers cracked an SF rival's tech model with a single word​

A research team from the tech giant got ChatGPT to spit out its private training data​

By Stephen CouncilDec 1, 2023



Demis Hassabis, the CEO and co-founder of DeepMind, attends an AI Safety Summit on Nov. 2, 2023, in Bletchley, England. DeepMind, an artificial intelligence research lab, was purchased by Google in 2014.

Toby Melville - WPA Pool/Getty Images

Just in time for ChatGPT to turn a year old, a group of researchers from Google published a paper showing how easy it is to break OpenAI’s buzzy technology.

The paper, published Tuesday, provides a look at how scientists at the forefront of artificial intelligence research — an extremely well-paid job, for some — are testing the limits of popular products in real time. Google and its AI lab, DeepMind, where the majority of the paper’s authors work, are in a race to turn scientific advancements into lucrative and useful products, before rivals like OpenAI and Meta get there first.

The study takes a look at “extraction,” which is an “adversarial” attempt to glean what data might have been used to train an AI tool. AI models “memorize examples from their training datasets, which can allow an attacker to extract (potentially private) information,” the researchers wrote. The privacy is key: If AI models are eventually trained on personal information, breaches of their training data could reveal bank logins, home addresses and more.

ChatGPT, the Google team added in a blog post announcing the paper, is “‘aligned’ to not spit out large amounts of training data. But, by developing an attack, we can do exactly this.” Alignment, in AI, refers to engineers’ attempts to guide the tech’s behavior. The researchers also noted that ChatGPT is a product that has been released to the market for public use, as opposed to previous production-phase AI models that have succumbed to extraction attempts.

The “attack” that worked was so simple, the researchers even called it “silly” in their blog post: They just asked ChatGPT to repeat the word “poem” forever.

They found that, after repeating “poem” hundreds of times, the chatbot would eventually “diverge,” or leave behind its standard dialogue style and start spitting out nonsensical phrases. When the researchers repeated the trick and looked at the chatbot’s output (after the many, many “poems”), they began to see content that was straight from ChatGPT’s training data. They had figured out “extraction,” on a cheap-to-use version of the world’s most famous AI chatbot, “ChatGPT-3.5-turbo.”

After running similar queries again and again, the researchers had used just $200 to get more than 10,000 examples of ChatGPT spitting out memorized training data, they wrote. This included verbatim paragraphs from novels, the personal information of dozens of people, snippets of research papers and “NSFW content” from dating sites, according to the paper.

404 Media, which first reported on the paper, found several of the passages online, including on CNN’s website, Goodreads, fan pages, blogs and even within comments sections.

The researchers wrote in their blog post, “As far as we can tell, no one has ever noticed that ChatGPT emits training data with such high frequency until this paper. So it’s worrying that language models can have latent vulnerabilities like this.”

“It’s also worrying that it’s very hard to distinguish between (a) actually safe and (b) appears safe but isn’t,” they added. Along with Google, the research team featured representatives from UC Berkeley, University of Washington, Cornell, Carnegie Mellon and ETH Zurich.

The researchers wrote in the paper that they told OpenAI about ChatGPT’s vulnerability on Aug. 30, giving the startup time to fix the issue before the team publicized its findings. But on Thursday afternoon, SFGATE was able to replicate the issue: When asked to repeat just the word “ripe” forever, the public and free version of ChatGPT eventually started spitting out other text, including quotes correctly attributed to Richard Bach and Toni Morrison.

OpenAI did not immediately respond to SFGATE’s request for comment. On Wednesday, the company officially welcomed Sam Altman back as CEO, after a dramatic ouster that consumed the startup a couple weeks ago.

 

[bnew]

 

[bnew]

 

[bnew]

The Robots Will Insider Trade

Also OpenAI’s board, kangaroo grazing and bank box-checking.

www.bloomberg.com

The Robots Will Insider Trade​

---

[2311.07590] Large Language Models can Strategically Deceive their Users when Put Under Pressure

Computer Science > Computation and Language​

[Submitted on 9 Nov 2023 (v1), last revised 27 Nov 2023 (this version, v2)]

Technical Report: Large Language Models can Strategically Deceive their Users when Put Under Pressure​

Jérémy Scheurer, Mikita Balesni, Marius Hobbhahn

We demonstrate a situation in which Large Language Models, trained to be helpful, harmless, and honest, can display misaligned behavior and strategically deceive their users about this behavior without being instructed to do so. Concretely, we deploy GPT-4 as an agent in a realistic, simulated environment, where it assumes the role of an autonomous stock trading agent. Within this environment, the model obtains an insider tip about a lucrative stock trade and acts upon it despite knowing that insider trading is disapproved of by company management. When reporting to its manager, the model consistently hides the genuine reasons behind its trading decision. We perform a brief investigation of how this behavior varies under changes to the setting, such as removing model access to a reasoning scratchpad, attempting to prevent the misaligned behavior by changing system instructions, changing the amount of pressure the model is under, varying the perceived risk of getting caught, and making other simple changes to the environment. To our knowledge, this is the first demonstration of Large Language Models trained to be helpful, harmless, and honest, strategically deceiving their users in a realistic situation without direct instructions or training for deception.

Subjects:	Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as:	arXiv:2311.07590 [cs.CL]
	(or arXiv:2311.07590v2 [cs.CL] for this version)
	https://doi.org/10.48550/arXiv.2311.07590 Focus to learn more

Submission history​

From: Jérémy Scheurer [view email]
[v1] Thu, 9 Nov 2023 17:12:44 UTC (812 KB)
[v2] Mon, 27 Nov 2023 15:17:49 UTC (805 KB)

https://arxiv.org/pdf/2311.07590

 

[bnew]

https://www.reuters.com/technology/generative-ai-stumbling-block-eu-legislation-talks-sources-2023-12-01/

Generative AI a stumbling block in EU legislation talks -sources​

By Supantha Mukherjee, Foo Yun Chee and Martin Coulter

December 1, 2023
4:13 PM EST
Updated 2 days ago

[1/2]Technology leaders attend a generative AI (Artificial Intelligence) meeting in San Francisco as the city is trying to position itself as the “AI capital of the world”, in California, U.S., June 29, 2023. REUTERS/Carlos Barria/File Photo Acquire Licensing Rights

STOCKHOLM/BRUSSELS/LONDON, Dec 1 (Reuters) - EU lawmakers cannot agree on how to regulate systems like ChatGPT, in a threat to landmark legislation aimed at keeping artificial intelligence (AI) in check, six sources told Reuters.

As negotiators meet on Friday for crucial discussions ahead of final talks scheduled for Dec. 6, 'foundation models', or generative AI, have become the main hurdle in talks over the European Union's proposed AI Act, said the sources, who declined to be identified because the discussions are confidential.

Foundation models like the one built by Microsoft (MSFT.O)-backed OpenAI are AI systems trained on large sets of data, with the ability to learn from new data to perform various tasks.

After two years of negotiations, the bill was approved by the European parliament in June. The draft AI rules now need to be agreed through meetings between representatives of the European Parliament, the Council and the European Commission.

Experts from EU countries will meet on Friday to thrash out their position on foundation models, access to source codes, fines and other topics while lawmakers from the European Parliament are also gathering to finalise their stance.

If they cannot agree, the act risks being shelved due to lack of time before European parliamentary elections next year.

While some experts and lawmakers have proposed a tiered approach for regulating foundation models, defined as those with more than 45 million users, others have said smaller models could be equally risky.

But the biggest challenge to getting an agreement has come from France, Germany and Italy, who favour letting makers of generativeAI models self-regulate instead of having hard rules.

In a meeting of the countries' economy ministers on Oct. 30 in Rome, France persuaded Italy and Germany to support a proposal, sources told Reuters.

Until then, negotiations had gone smoothly, with lawmakers making compromises across several other conflict areas such as regulating high-risk AI, sources said.

SELF-REGULATION?​

European parliamentarians, EU Commissioner Thierry Breton and scores of AI researchers have criticised self-regulation.

In an open letter this week, researchers such as Geoffrey Hinton warned self-regulation is "likely to dramatically fall short of the standards required for foundation model safety".

France-based AI company Mistral and Germany's Aleph Alpha have criticised the tiered approach to regulating foundation models, winning support from their respective countries.

A source close to Mistral said the company favours hard rules for products, not the technology on which it is built.

"Though the concerned stakeholders are working their best to keep negotiations on track, the growing legal uncertainty is unhelpful to European industries,” said Kirsten Rulf, a Partner and Associate Director at Boston Consulting Group.

“European businesses would like to plan for next year, and many want to see some kind of certainty around the EU AI Act going into 2024,” she added.

Other pending issues in the talks include definition of AI, fundamental rights impact assessment, law enforcement exceptions and national security exceptions, sources told Reuters.

Lawmakers have also been divided over the use of AI systems by law enforcement agencies for biometric identification of individuals in publicly accessible spaces and could not agree on several of these topics in a meeting on Nov. 29, sources said.

Spain, which holds the EU presidency until the end of the year, has proposed compromises in a bid to speed up the process.

If a deal does not happen in December, the next presidency Belgium will have a couple of months to one before it is likely shelved ahead of European elections.

"Had you asked me six or seven weeks ago, I would have said we are seeing compromises emerging on all the key issues," said Mark Brakel, director of policy at the Future of Life Institute, a nonprofit aimed at reducing risks from advanced AI.

"This has now become a lot harder," he said.

Reporting by Supantha Mukherjee in Stockholm; Editing by Josephine Mason and Alexander Smith, Kirsten Donova

Our Standards: The Thomson Reuters Trust Principles.