뉴스 | 🇺🇸 미국 | "SIMPLY" | 3페이지

This sponsored article is brought to you by Ampace. As AI workloads grow to gigascale levels, the global data center industry has hit a hidden physical wall. The real bottleneck is no longer just the thermal limit of the chip or the capacity of the cooling system — it is the dynamic resilience of the power chain. Modern AI computing clusters, driven by massive GPU clusters, generate high-frequency, abrupt, and synchronized spikey pulse loads. As rack densities soar beyond 100 kW, these fluctuations are amplified into a “power paradox”: while the digital logic of AI is moving faster than ever, the physical infrastructure supporting it remains tethered to legacy response capabilities. The power usage of these gigascale sites and their drastic, high frequency, abrupt load surges from the AI GPU clusters can trigger transient voltage events and frequency instability, risking the entire local grid. The grid itself is not robust enough to support these loads. This leads to the infrastructure gap: The utility is not robust enough and traditional backup sources, such as diesel generators and gas turbines, simply cannot react to millisecond-level power spikes in output. This will often force operators into a cycle of costly infrastructure over sizing just to buffer the volatility. AI infrastructure requires energy systems capable of instantaneous response while safeguarding continuity and reliability. The industry has explored various mitigations — from rack-level BBUs to 800V DC architectures — yet the mature, high volume, traditional UPS system remains the most viable and scalable foundation for gigawatt-level facilities. Consequently, the UPS-integrated battery system has emerged as the critical “physical buffer” to neutralize these pulses at the source. At Data Center World 2026 in Washington, D.C., Ampace led a pivotal technical dialogue with Eaton during the session “Powering Giga-scale AI.” Their exchange unveiled a fundamental paradigm shift: To bridge the AI power gap, energy storage must evolve from a passive insurance policy into an active, high-speed stabilizer. By aligning Ampace’s semi-solid-state battery innovation with Eaton’s proven system intelligence, we are moving beyond simple backup to solve the physical paradox of the AI era. To move beyond simple backup and solve the physical paradox of the AI era, Ampace is aligning its semi-solid-state battery innovation with Eaton’s proven system intelligence.Ampace The “Shock Absorber” physics: semi-solid chemistry for AI pulses Conventional power systems were designed for steady-state loads, not the rapid heartbeat of a massive AI GPU cluster. When thousands of GPUs synchronize their computing cycles, they generate high-frequency, abrupt pulse loads that can lead to voltage sags, frequency oscillations, and potential interruptions of critical AI training. Ampace’s PU Series semi-solid and low-electrolyte cells address this challenge by acting as high-speed “shock absorbers.” Leveraging ultra-low internal resistance (DCR) and high cycle capability, these batteries neutralize millisecond-level power spikes at the source, stabilizing the local power loop before disturbances propagate upstream to the grid or on-site generators. These high-rate cells enable 100 kW+ racks to maintain peak performance without transmitting instability across the power chain. This capability aligns closely with Eaton’s matured UPS architectures, such as double-conversion topologies and advanced power electronics upgrades, which have long prioritized rapid load responsiveness and high system stability. Together, these approaches embody a shared industry philosophy: AI infrastructure requires energy systems capable of instantaneous response while safeguarding continuity and reliability. Ampace’s semi-solid state chemistry minimizes liquid electrolyte, greatly reducing the risk of leakage and thermal runaway under continuous AI high-load conditions.Ampace Algorithmic intelligence: synchronizing energy and control Hardware alone cannot solve the AI power paradox; the system also requires intelligent coordination between energy storage and power management. Sophisticated battery management systems (BMS) like Ampace’s high-precision design track state-of-charge (SOC) with high-speed sampling, even during rapid, shallow cycling typical in AI workloads. Complementary algorithmic approaches in modern UPS platforms — such as ramp-rate control and average power management — effectively suppress sub-synchronous oscillations and optimize load smoothing. In large-scale AI training environments, where thousands of GPUs can trigger millisecond-level power pulses, these intelligent layers ensure that batteries buffer high-frequency fluctuations without compromising the mandatory emergency backup reserves. By transforming energy storage from passive “standby insurance” into active, schedulable assets, the system simultaneously safeguards continuous AI training and maintains the long-term health of the data center infrastructure. In practical terms, this means that even during peak compute bursts, the infrastructure remains stable, training cycles continue uninterrupted, and operators avoid costly oversizing or grid stress. Eaton’s dual-layer algorithms serve as a valuable benchmark in this space, demonstrating how advanced control logic can achieve similar objectives, reinforcing Ampace’s approach and philosophy within the broader data center power ecosystem. Economic scalability: optimizing AI infrastructure efficiently One of the largest costs in deploying AI infrastructure is “oversizing”: procuring transformers, generators, and UPS systems to handle brief peak spikes. This traditional approach inflates the Total Cost of Ownership (TCO) and leads to wasted capital on underutilized hardware. Ampace’s turn-key cabinet design developed by its independent R&D is engineered for seamless compatibility with mature, high volume UPS systems. By leveraging Eaton’s double-conversion UPS topologies alongside intelligent ramp-rate and average power management algorithms, AI data centers can scale dynamically without requiring costly infrastructure redesigns. This approach allows the UPS and batteries to act as active load-shapers, smoothing AI-driven pulses while strictly maintaining mandatory emergency backup capacity. By utilizing energy storage as an active, schedulable asset, operators can right-size their infrastructure, avoid unnecessary grid upgrades, and deploy gigascale AI clusters with unprecedented efficiency. Safety First: Protecting AI Infrastructure While Enabling Innovation In high-density AI facilities, safety is non-negotiable. Ampace’s semi-solid state chemistry minimizes liquid electrolyte, greatly reducing the risk of leakage and thermal runaway under continuous AI high-load conditions. Ampace’s turn-key cabinet design developed by its independent R&D is engineered for seamless compatibility with mature, high volume UPS systems. Ampace At the same time, Eaton’s UPS design emphasizes system-level energy scheduling that never sacrifices mandatory emergency backup reserves, ensuring thermal safety and uninterrupted operation. This “safety-first” approach ensures that infrastructure can sustain aggressive performance targets without compromising the physical integrity of the facility. Coupled with over a decade of proven high-cycle life operation and design under shallow pulse conditions, these systems can extend operational lifespan, reduce replacement requirements, and provide operators with confidence that safety and reliability remain uncompromised as compute density continues to grow. To remain the scalable backbone of AI data centers As AI computing scales over the next two to three years, the industry will face stricter grid requirements and even more demanding pulse load characteristics. This evolution demands a forward-looking design philosophy that harmonizes UPS, battery, and grid compatibility. Ampace views current low-electrolyte semi-solid technologies as the optimal transitional step toward a fully solid-state future — one that promises ultimate safety and performance. Ampace remains committed to this long-term technological roadmap. We view current low-electrolyte semi-solid technologies as the optimal transitional step toward a fully solid-state future — one that promises ultimate safety and performance. Whether through rack-level BBU, integrated UPS systems, or containerized storage, the universal core of the AI era remains constant: high-speed response, long shallow-cycle life, and refined energy management. By engaging in deep technical exchanges with Eaton and leading energy innovators, Ampace ensures that its solutions not only meet today’s AI pulse challenges but also harmonize with broader infrastructure strategies and shared industry best practices. Ultimately, as traditional diesel generators gradually give way to diversified alternatives, the integrated UPS-plus-energy-storage system will become the fundamental infrastructure standard. The dialogue has just begun. Ampace will continue to engage in strategic exchanges with global industrial automation leaders and digital energy pioneers, co-authoring the playbook for a safer, more efficient, and more resilient AI-ready world.

IEEE Spectrum

중도 성향

IT/기술

With $1 Cyberattacks on the Rise, Durable Defenses Pay Off

Transforming a newly discovered software vulnerability into a cyberattack used to take months. Today—as the recent headlines over Anthropic’s Project Glasswing have shown—generative AI can do the job in minutes, often for less than a dollar of cloud-computing time. But while large language models present a real cyberthreat, they also provide an opportunity to reinforce cyberdefenses. Anthropic reports its Claude Mythos preview model has already helped defenders preemptively discover over a thousand zero-day vulnerabilities, including flaws in every major operating system and web browser, with Anthropic coordinating disclosure and its efforts to patch the revealed flaws. It is not yet clear whether AI-driven bug finding will ultimately favor attackers or defenders. But to understand how defenders can increase their odds, and perhaps hold the advantage, it helps to look at an earlier wave of automated vulnerability discovery. In the early 2010s, a new category of software appeared that could attack programs with millions of random, malformed inputs—a proverbial monkey at a typewriter, tapping on the keys until it finds a vulnerability. When such “fuzzers” like American Fuzzy Lop (AFL) hit the scene, they found critical flaws in every major browser and operating system. The security community’s response was instructive. Rather than panic, organizations industrialized the defense. For instance, Google built a system called OSS-Fuzz that runs fuzzers continuously, around the clock, on thousands of software projects. So software providers could catch bugs before they shipped, not after attackers found them. The expectation is that AI-driven vulnerability discovery will follow the same arc. Organizations will integrate the tools into standard development practice, run them continuously, and establish a new baseline for security. But the analogy has a limit. Fuzzing requires significant technical expertise to set up and operate. It was a tool for specialists. An LLM, meanwhile, finds vulnerabilities with just a prompt—resulting in a troubling asymmetry. Attackers no longer need to be technically sophisticated to exploit code, while robust defenses still require engineers to read, evaluate, and act on what the AI models surface. The human cost of finding and exploiting bugs may approach zero, but fixing them won’t. Is AI Better at Finding Bugs Than Fixing Them? In the opening to his book Engineering Security (2014), Peter Gutmann observed that “a great many of today’s security technologies are ‘secure’ only because no one has ever bothered to look at them.” That observation was made before AI made looking for bugs dramatically cheaper. Most present-day code—including the open source infrastructure that commercial software depends on—is maintained by small teams, part-time contributors, or individual volunteers with no dedicated security resources. A bug in any open source project can have significant downstream impact, too. In 2021, a critical vulnerability in Log4j—a logging library maintained by a handful of volunteers—exposed hundreds of millions of devices. Log4j’s widespread use meant that a vulnerability in a single volunteer-maintained library became one of the most widespread software vulnerabilities ever recorded. The popular code library is just one example of the broader problem of critical software dependencies that have never been seriously audited. For better or worse, AI-driven vulnerability discovery will likely perform a lot of auditing, at low cost and at scale. An attacker targeting an under-resourced project requires little manual effort. AI tools can scan an unaudited codebase, identify critical vulnerabilities, and assist in building a working exploit with minimal human expertise. Research on LLM-assisted exploit generation has shown that capable models can autonomously and rapidly exploit cyber weaknesses, compressing the time between disclosure of the bug and working exploit of that bug from weeks down to mere hours. Generative AI-based attacks launched from cloud servers operate staggeringly cheaply as well. In August 2025, researchers at NYU’s Tandon School of Engineering demonstrated that an LLM-based system could autonomously complete the major phases of a ransomware campaign for some $0.70 per run, with no human intervention. And the attacker’s job ends there. The defender’s job, on the other hand, is only getting underway. While an AI tool can find vulnerabilities and potentially assist with bug triaging, a dedicated security engineer still has to review any potential patches, evaluate the AI’s analysis of the root cause, and understand the bug well enough to approve and deploy a fully functional fix without breaking anything. For a small team maintaining a widely-depended-upon library in their spare time, that remediation burden may be difficult to manage even if the discovery cost drops to zero. Why AI Guardrails and Automated Patching Aren’t the Answer The natural policy response to the problem is to go after AI at the source: holding AI companies responsible for spotting misuse, putting guardrails in their products, and pulling the plug on anyone using LLMs to mount cyberattacks. There is evidence that pre-emptive defenses like this have some effect. Anthropic has published data showing that automated misuse detection can derail some cyberattacks. However, blocking a few bad actors does not make for a satisfying and comprehensive solution. At a root level, there are two reasons why policy does not solve the whole problem. The first is technical. LLMs judge whether a request is malicious by reading the request itself. But a sufficiently creative prompt can frame any harmful action as a legitimate one. Security researchers know this as the problem of the persuasive prompt injection. Consider, for example, the difference between “Attack website A to steal users’ credit card info” and “I am a security researcher and would like secure website A. Run a simulation there to see if it’s possible to steal users’ credit card info.” No one’s yet discovered how to root out the source of subtle cyberattacks, like in the latter example, with 100 percent accuracy. The second reason is jurisdictional. Any regulation confined to U.S.-based providers (or that of any other single country or region) still leaves the problem largely unsolved worldwide. Strong, open-source LLMs are already available anywhere the internet reaches. A policy aimed at handful of American technology companies is not a comprehensive defense. Another tempting fix is to automate the defensive side entirely—let AI autonomously identify, patch, and deploy fixes without waiting for an overworked volunteer maintainer to review them. Tools like GitHub Copilot Autofix generate patches for flagged vulnerabilities directly with proposed code changes. Several open-source security initiatives are also experimenting with autonomous AI maintainers for under-resourced projects. It is becoming much easier to have the same AI system find bugs, generate a patch, and update the code with no human intervention. But LLM-generated patches can be unreliable in ways that are difficult to detect. For example, even if they pass muster with popular code-testing software suites, they may still introduce subtle logic errors. LLM-generated code, even from the most powerful generative AI models out there, is still subject to a range of cyber-vulnerabilities. A coding agent with write access to a repository and no human in the loop is, in so many words, an easy target. Misleading bug reports, malicious instructions hidden in project files, or untrusted code pulled in from outside the project can turn an automated AI codebase maintainer into a cyber-vulnerability generator. Guardrails and automated patching are useful tools, but they share a common limitation. Both are ad hoc and incomplete. Neither addresses the deeper question of whether the software was built securely from the start. The more lasting solution is to prevent vulnerabilities from being introduced at all. No matter how deeply an AI system can inspect a project, it cannot find flaws that don’t exist. Memory-Safe Code Creates More Robust Defenses The most accessible starting point is the adoption of memory-safe languages. Simply by changing the programming language their coders use, organizations can have a large positive impact on their security. Both Google and Microsoft have found that roughly 70 percent of serious security flaws come down to the ways in which software manages memory. Languages like C and C++ leave every memory decision to the developer. And when something slips, even briefly, attackers can exploit that gap to run their own code, siphon data, or bring systems down. Languages like Rust go further; they make the most dangerous class of memory errors structurally impossible, not just harder to make. Memory-safe languages address the problem at the source, but legacy codebases written in C and C++ will remain a reality for decades. Software sandboxing techniques complement memory-safe languages by addressing what they cannot—containing the blast radius of vulnerabilities that do exist. Tools like WebAssembly and RLBox already demonstrate this in practice in web browsers and cloud service providers like Fastly and Cloudflare. However, while sandboxes dramatically raise the bar for attackers, they are only as strong as their implementation. Moreover, Anthropic reports that Claude Mythos has demonstrated that it can breach software sandboxes. For the most security-critical components, where implementation complexity is highest and the cost of failure greatest, a stronger guarantee still is available. Formal verification proves, mathematically, that certain bugs cannot exist. It treats code like a mathematical theorem. Instead of testing whether bugs appear, it proves that specific categories of flaw cannot exist under any conditions. AWS, Cloudflare, and Google already use formal verification to protect their most sensitive infrastructure—cryptographic code, network protocols, and storage systems where failure isn’t an option. Tools like Flux now bring that same rigor to everyday production Rust code, without requiring a dedicated team of specialists. That matters when your attacker is a powerful generative-AI system that can rapidly scan millions of lines of code for weaknesses. Formally verified code doesn’t just put up some fences and firewalls—it provably has no weaknesses to find. The defenses described above are asymmetric. Code written in memory-safe languages—separated by strong sandboxing boundaries and selectively formally verified—presents a smaller and much more constrained target. When applied correctly, these techniques can prevent LLM-powered exploitation, regardless of how capable an attacker’s bug-scanning tools become. Generative AI can support this more foundational shift by accelerating the translation of legacy code into safer languages like Rust, and making formal verification more practical at every stage. Which helps engineers write specifications, generate proofs, and keep those proofs current as code evolves. For organizations, the lasting solution is not just better scanning but stronger foundations: memory-safe languages where possible, sandboxing where not, and formal verification where the cost of being wrong is highest. For researchers, the bottleneck is making those foundations practical—and using generative AI to accelerate the migration. But instead of automated, ad hoc vulnerability patching, generative AI in this mode of defense can help translate legacy code to memory-safe alternatives. It also assists in verification proofs and lowers the expertise barrier to a safer and less vulnerable codebase. The latest wave of smarter AI bug scanners can still be useful for cyberdefense—not just as another overhyped AI threat. But AI bug scanners treat the symptom, not the cause. The lasting solution is software that doesn’t produce vulnerabilities in the first place.

IEEE Spectrum

중도 성향

IT/기술

1

Better Hardware Could Turn Zeros into AI Heroes

When it comes to AI models, size matters. Even though some artificial-intelligence experts warn that scaling up large language models (LLMs) is hitting diminishing performance returns, companies are still coming out with ever larger AI tools. Meta’s latest Llama release had a staggering 2 trillion parameters that define the model. As models grow in size, their capabilities increase. But so do the energy demands and the time it takes to run the models, which increases their carbon footprint. To mitigate these issues, people have turned to smaller, less capable models and using lower-precision numbers whenever possible for the model parameters. But there is another path that may retain a staggeringly large model’s high performance while reducing the time it takes to run an energy footprint. This approach involves befriending the zeros inside large AI models. For many models, most of the parameters—the weights and activations—are actually zero, or so close to zero that they could be treated as such without losing accuracy. This quality is known as sparsity. Sparsity offers a significant opportunity for computational savings: Instead of wasting time and energy adding or multiplying zeros, these calculations could simply be skipped; rather than storing lots of zeros in memory, one need only store the nonzero parameters. Unfortunately, today’s popular hardware, like multicore CPUs and GPUs, do not naturally take full advantage of sparsity. To fully leverage sparsity, researchers and engineers need to rethink and re-architect each piece of the design stack, including the hardware, low-level firmware, and application software. In our research group at Stanford University, we have developed the first (to our knowledge) piece of hardware that’s capable of calculating all kinds of sparse and traditional workloads efficiently. The energy savings varied widely over the workloads, but on average our chip consumed one-seventieth the energy of a CPU, and performed the computation on average eight times as fast. To do this, we had to engineer the hardware, low-level firmware, and software from the ground up to take advantage of sparsity. We hope this is just the beginning of hardware and model development that will allow for more energy-efficient AI. What is sparsity? Neural networks, and the data that feeds into them, are represented as arrays of numbers. These arrays can be one-dimensional (vectors), two-dimensional (matrices), or more (tensors). A sparse vector, matrix, or tensor has mostly zero elements. The level of sparsity varies, but when zeroes make up more than 50 percent of any type of array, it can stand to benefit from sparsity-specific computational methods. In contrast, an object that is not sparse—that is, it has few zeros compared with the total number of elements—is called dense. Sparsity can be naturally present, or it can be induced. For example, a social-network graph will be naturally sparse. Imagine a graph where each node (point) represents a person, and each edge (a line segment connecting the points) represents a friendship. Since most people are not friends with one another, a matrix representing all possible edges will be mostly zeros. Other popular applications of AI, such as other forms of graph learning and recommendation models, contain naturally occurring sparsity as well. Beyond naturally occurring sparsity, sparsity can also be induced within an AI model in several ways. Two years ago, a team at Cerebras showed that one can set up to 70 to 80 percent of parameters in an LLM to zero without losing any accuracy. Cerebras demonstrated these results specifically on Meta’s open-source Llama 7B model, but the ideas extend to other LLM models like ChatGPT and Claude. The case for sparsity Sparse computation’s efficiency stems from two fundamental properties: the ability to compress away zeros and the convenient mathematical properties of zeros. Both the algorithms used in sparse computation and the hardware dedicated to them leverage these two basic ideas. First, sparse data can be compressed, making it more memory efficient to store “sparsely”—that is, in something called a sparse data type. Compression also makes it more energy efficient to move data when dealing with large amounts of it. This is best understood by an example. Take a four-by-four matrix with three nonzero elements. Traditionally, this matrix would be stored in memory as is, taking up 16 spaces. This matrix can also be compressed into a sparse data type, getting rid of the zeros and saving only the nonzero elements. In our example, this results in 13 memory spaces as opposed to 16 for the dense, uncompressed version. These savings in memory increase with increased sparsity and matrix size. In addition to the actual data values, compressed data also requires metadata. The row and column locations of the nonzero elements also must be stored. This is usually thought of as a “fibertree”: The row labels containing nonzero elements are listed and linked to the column labels of the nonzero elements, which are then linked to the values stored in those elements. In memory, things get a bit more complicated still: The row and column labels for each nonzero value must be stored as well as the “segments” that indicate how many such labels to expect, so the metadata and data can be clearly delineated from one another. In a dense, noncompressed matrix data type, values can be accessed either one at a time or in parallel, and their locations can be calculated directly with a simple equation. However, accessing values in sparse, compressed data requires looking up the coordinates of the row index and using that information to “indirectly” look up the coordinates of the column index before finally reaching the value. Depending on the actual locations of the sparse data values, these indirect lookups can be extremely random, making the computation data-dependent and requiring the allocation of memory lookups on the fly. Second, two mathematical properties of zero let software and hardware skip a lot of computation. Multiplying any number by zero will result in a zero, so there’s no need to actually do the multiplication. Adding zero to any number will always return that number, so there’s no need to do the addition either. In matrix-vector multiplication, one of the most common operations in AI workloads, all computations except those involving two nonzero elements can simply be skipped. Take, for example, the four-by-four matrix from the previous example and a vector of four numbers. In dense computation, each element of the vector must be multiplied by the corresponding element in each row and then added together to compute the final vector. In this case, that would take 16 multiplication operations and 16 additions (or four accumulations). In sparse computation, only the nonzero elements of the vector need be considered. For each nonzero vector element, indirect lookup can be used to find any corresponding nonzero matrix element, and only those need to be multiplied and added. In the example shown here, only two multiplication steps will be performed, instead of 16. The trouble with GPUs and CPUs Unfortunately, modern hardware is not well suited to accelerating sparse computation. For example, say we want to perform a matrix-vector multiplication. In the simplest case, in a single CPU core, each element in the vector would be multiplied sequentially and then written to memory. This is slow, because we can do only one multiplication at a time. So instead people use CPUs with vector support or GPUs. With this hardware, all elements would be multiplied in parallel, greatly speeding up the application. Now, imagine that both the matrix and vector contain extremely sparse data. The vectorized CPU and GPU would spend most of their efforts multiplying by zero, performing completely ineffectual computations. Newer generations of GPUs are capable of taking some advantage of sparsity in their hardware, but only a particular kind, called structured sparsity. Structured sparsity assumes that two out of every four adjacent parameters are zero. However, some models benefit more from unstructured sparsity—the ability for any parameter (weight or activation) to be zero and compressed away, regardless of where it is and what it is adjacent to. GPUs can run unstructured sparse computation in software, for example, through the use of the cuSparse GPU library. However, the support for sparse computations is often limited, and the GPU hardware gets underutilized, wasting energy-intensive computations on overhead. Petra Péterffy When doing sparse computations in software, modern CPUs may be a better alternative to GPU computation, because they are designed to be more flexible. Yet, sparse computations on the CPU are often bottlenecked by the indirect lookups used to find nonzero data. CPUs are designed to “prefetch” data based on what they expect they’ll need from memory, but for randomly sparse data, that process often fails to pull in the right stuff from memory. When that happens, the CPU must waste cycles calling for the right data. Apple was the first to speed up these indirect lookups by supporting a method called an array-of-pointers access pattern in the prefetcher of their A14 and M1 chips. Although innovations in prefetching make Apple CPUs more competitive for sparse computation, CPU architectures still have fundamental overheads that a dedicated sparse computing architecture would not, because they need to handle general-purpose computation. Other companies have been developing hardware that accelerates sparse machine learning as well. These include Cerebras’s Wafer Scale Engine and Meta’s Training and Inference Accelerator (MTIA). The Wafer Scale Engine, and its corresponding sparse programming framework, have shown incredibly sparse results of up to 70 percent sparsity on LLMs. However, the company’s hardware and software solutions support only weight sparsity, not activation sparsity, which is important for many applications. The second version of the MTIA claims a sevenfold sparse compute performance boost over the MTIA v1. However, the only publicly available information regarding sparsity support in the MTIA v2 is for matrix multiplication, not for vectors or tensors. Although matrix multiplications take up the majority of computation time in most modern ML models, it’s important to have sparsity support for other parts of the process. To avoid switching back and forth between sparse and dense data types, all of the operations should be sparse. Onyx Instead of these halfway solutions, our team at Stanford has developed a hardware accelerator, Onyx, that can take advantage of sparsity from the ground up, whether it’s structured or unstructured. Onyx is the first programmable accelerator to support both sparse and dense computation; it’s capable of accelerating key operations in both domains. To understand Onyx, it is useful to know what a coarse-grained reconfigurable array (CGRA) is and how it compares with more familiar hardware, like CPUs and field-programmable gate arrays (FPGAs). CPUs, CGRAs, and FPGAs represent a trade-off between efficiency and flexibility. Each individual logic unit of a CPU is designed for a specific function that it performs efficiently. On the other hand, since each individual bit of an FPGA is configurable, these arrays are extremely flexible, but very inefficient. The goal of CGRAs is to achieve the flexibility of FPGAs with the efficiency of CPUs. CGRAs are composed of efficient and configurable units, typically memory and compute, that are specialized for a particular application domain. This is the key benefit of this type of array: Programmers can reconfigure the internals of a CGRA at a high level, making it more efficient than an FPGA but more flexible than a CPU. The Onyx chip, built on a coarse-grained reconfigurable array (CGRA), is the first (to our knowledge) to support both sparse and dense computations. Olivia Hsu Onyx is composed of flexible, programmable processing element (PE) tiles and memory (MEM) tiles. The memory tiles store compressed matrices and other data formats. The processing element tiles operate on compressed matrices, eliminating all unnecessary and ineffectual computation. The Onyx compiler handles conversion from software instructions to CGRA configuration. First, the input expression—for instance, a sparse vector multiplication—is translated into a graph of abstract memory and compute nodes. In this example, there are memories for the input vectors and output vectors, a compute node for finding the intersection between nonzero elements, and a compute node for the multiplication. The compiler figures out how to map the abstract memory and compute nodes onto MEMs and PEs on the CGRA, and then how to route them together so that they can transfer data between them. Finally, the compiler produces the instruction set needed to configure the CGRA for the desired purpose. Since Onyx is programmable, engineers can map many different operations, such as vector-vector element multiplication, or the key tasks in AI, like matrix-vector or matrix-matrix multiplication, onto the accelerator. We evaluated the efficiency gains of our hardware by looking at the product of energy used and the time it took to compute, called the energy-delay product (EDP). This metric captures the trade-off of speed and energy. Minimizing just energy would lead to very slow devices, and minimizing speed would lead to high-area, high-power devices. Onyx achieves up to 565 times as much energy-delay product over CPUs (we used a 12-core Intel Xeon CPU) that utilize dedicated sparse libraries. Onyx can also be configured to accelerate regular, dense applications, similar to the way a GPU or TPU would. If the computation is sparse, Onyx is configured to use sparse primitives, and if the computation is dense, Onyx is reconfigured to take advantage of parallelism, similar to how GPUs function. This architecture is a step toward a single system that can accelerate both sparse and dense computations on the same silicon. Just as important, Onyx enables new algorithmic thinking. Sparse acceleration hardware will not only make AI more performance- and energy efficient but also enable researchers and engineers to explore new algorithms that have the potential to dramatically improve AI. The future with sparsity Our team is already working on next-generation chips built off of Onyx. Beyond matrix multiplication operations, machine learning models perform other types of math, like nonlinear layers, normalization, the softmax function, and more. We are adding support for the full range of computations on our next-gen accelerator and within the compiler. Since sparse machine learning models may have both sparse and dense layers, we are also working on integrating the dense and sparse accelerator architecture more efficiently on the chip, allowing for fast transformation between the different data types. We’re also looking at ways to manage memory constraints by breaking up the sparse data more effectively so we can run computations on several sparse accelerator chips. We are also working on systems that can predict the performance of accelerators such as ours, which will help in designing better hardware for sparse AI. Longer term, we’re interested in seeing whether high degrees of sparsity throughout AI computation will catch on with more model types, and whether sparse accelerators become adopted at a larger scale. Building the hardware to unstructured sparsity and optimally take advantage of zeros is just the beginning. With this hardware in hand, AI researchers and engineers will have the opportunity to explore new models and algorithms that leverage sparsity in novel and creative ways. We see this as a crucial research area for managing the ever-increasing runtime, costs, and environmental impact of AI.

IEEE Spectrum

중도 성향

IT/기술

Andrew Ng: Unbiggen AI

Andrew Ng has serious street cred in artificial intelligence. He pioneered the use of graphics processing units (GPUs) to train deep learning models in the late 2000s with his students at Stanford University, cofounded Google Brain in 2011, and then served for three years as chief scientist for Baidu, where he helped build the Chinese tech giant’s AI group. So when he says he has identified the next big shift in artificial intelligence, people listen. And that’s what he told IEEE Spectrum in an exclusive Q&A. Ng’s current efforts are focused on his company Landing AI, which built a platform called LandingLens to help manufacturers improve visual inspection with computer vision. He has also become something of an evangelist for what he calls the data-centric AI movement, which he says can yield “small data” solutions to big issues in AI, including model efficiency, accuracy, and bias. Andrew Ng on... What’s next for really big models The career advice he didn’t listen to Defining the data-centric AI movement Synthetic data Why Landing AI asks its customers to do the work The great advances in deep learning over the past decade or so have been powered by ever-bigger models crunching ever-bigger amounts of data. Some people argue that that’s an unsustainable trajectory. Do you agree that it can’t go on that way? Andrew Ng: This is a big question. We’ve seen foundation models in NLP [natural language processing]. I’m excited about NLP models getting even bigger, and also about the potential of building foundation models in computer vision. I think there’s lots of signal to still be exploited in video: We have not been able to build foundation models yet for video because of compute bandwidth and the cost of processing video, as opposed to tokenized text. So I think that this engine of scaling up deep learning algorithms, which has been running for something like 15 years now, still has steam in it. Having said that, it only applies to certain problems, and there’s a set of other problems that need small data solutions. When you say you want a foundation model for computer vision, what do you mean by that? Ng: This is a term coined by Percy Liang and some of my friends at Stanford to refer to very large models, trained on very large data sets, that can be tuned for specific applications. For example, GPT-3 is an example of a foundation model [for NLP]. Foundation models offer a lot of promise as a new paradigm in developing machine learning applications, but also challenges in terms of making sure that they’re reasonably fair and free from bias, especially if many of us will be building on top of them. What needs to happen for someone to build a foundation model for video? Ng: I think there is a scalability problem. The compute power needed to process the large volume of images for video is significant, and I think that’s why foundation models have arisen first in NLP. Many researchers are working on this, and I think we’re seeing early signs of such models being developed in computer vision. But I’m confident that if a semiconductor maker gave us 10 times more processor power, we could easily find 10 times more video to build such models for vision. Having said that, a lot of what’s happened over the past decade is that deep learning has happened in consumer-facing companies that have large user bases, sometimes billions of users, and therefore very large data sets. While that paradigm of machine learning has driven a lot of economic value in consumer software, I find that that recipe of scale doesn’t work for other industries. Back to top It’s funny to hear you say that, because your early work was at a consumer-facing company with millions of users. Ng: Over a decade ago, when I proposed starting the Google Brain project to use Google’s compute infrastructure to build very large neural networks, it was a controversial step. One very senior person pulled me aside and warned me that starting Google Brain would be bad for my career. I think he felt that the action couldn’t just be in scaling up, and that I should instead focus on architecture innovation. “In many industries where giant data sets simply don’t exist, I think the focus has to shift from big data to good data. Having 50 thoughtfully engineered examples can be sufficient to explain to the neural network what you want it to learn.” —Andrew Ng, CEO & Founder, Landing AI I remember when my students and I published the first NeurIPS workshop paper advocating using CUDA, a platform for processing on GPUs, for deep learning—a different senior person in AI sat me down and said, “CUDA is really complicated to program. As a programming paradigm, this seems like too much work.” I did manage to convince him; the other person I did not convince. I expect they’re both convinced now. Ng: I think so, yes. Over the past year as I’ve been speaking to people about the data-centric AI movement, I’ve been getting flashbacks to when I was speaking to people about deep learning and scalability 10 or 15 years ago. In the past year, I’ve been getting the same mix of “there’s nothing new here” and “this seems like the wrong direction.” Back to top How do you define data-centric AI, and why do you consider it a movement? Ng: Data-centric AI is the discipline of systematically engineering the data needed to successfully build an AI system. For an AI system, you have to implement some algorithm, say a neural network, in code and then train it on your data set. The dominant paradigm over the last decade was to download the data set while you focus on improving the code. Thanks to that paradigm, over the last decade deep learning networks have improved significantly, to the point where for a lot of applications the code—the neural network architecture—is basically a solved problem. So for many practical applications, it’s now more productive to hold the neural network architecture fixed, and instead find ways to improve the data. When I started speaking about this, there were many practitioners who, completely appropriately, raised their hands and said, “Yes, we’ve been doing this for 20 years.” This is the time to take the things that some individuals have been doing intuitively and make it a systematic engineering discipline. The data-centric AI movement is much bigger than one company or group of researchers. My collaborators and I organized a data-centric AI workshop at NeurIPS, and I was really delighted at the number of authors and presenters that showed up. You often talk about companies or institutions that have only a small amount of data to work with. How can data-centric AI help them? Ng: You hear a lot about vision systems built with millions of images—I once built a face recognition system using 350 million images. Architectures built for hundreds of millions of images don’t work with only 50 images. But it turns out, if you have 50 really good examples, you can build something valuable, like a defect-inspection system. In many industries where giant data sets simply don’t exist, I think the focus has to shift from big data to good data. Having 50 thoughtfully engineered examples can be sufficient to explain to the neural network what you want it to learn. When you talk about training a model with just 50 images, does that really mean you’re taking an existing model that was trained on a very large data set and fine-tuning it? Or do you mean a brand new model that’s designed to learn only from that small data set? Ng: Let me describe what Landing AI does. When doing visual inspection for manufacturers, we often use our own flavor of RetinaNet. It is a pretrained model. Having said that, the pretraining is a small piece of the puzzle. What’s a bigger piece of the puzzle is providing tools that enable the manufacturer to pick the right set of images [to use for fine-tuning] and label them in a consistent way. There’s a very practical problem we’ve seen spanning vision, NLP, and speech, where even human annotators don’t agree on the appropriate label. For big data applications, the common response has been: If the data is noisy, let’s just get a lot of data and the algorithm will average over it. But if you can develop tools that flag where the data’s inconsistent and give you a very targeted way to improve the consistency of the data, that turns out to be a more efficient way to get a high-performing system. “Collecting more data often helps, but if you try to collect more data for everything, that can be a very expensive activity.” —Andrew Ng For example, if you have 10,000 images where 30 images are of one class, and those 30 images are labeled inconsistently, one of the things we do is build tools to draw your attention to the subset of data that’s inconsistent. So you can very quickly relabel those images to be more consistent, and this leads to improvement in performance. Could this focus on high-quality data help with bias in data sets? If you’re able to curate the data more before training? Ng: Very much so. Many researchers have pointed out that biased data is one factor among many leading to biased systems. There have been many thoughtful efforts to engineer the data. At the NeurIPS workshop, Olga Russakovsky gave a really nice talk on this. At the main NeurIPS conference, I also really enjoyed Mary Gray’s presentation, which touched on how data-centric AI is one piece of the solution, but not the entire solution. New tools like Datasheets for Datasets also seem like an important piece of the puzzle. One of the powerful tools that data-centric AI gives us is the ability to engineer a subset of the data. Imagine training a machine-learning system and finding that its performance is okay for most of the data set, but its performance is biased for just a subset of the data. If you try to change the whole neural network architecture to improve the performance on just that subset, it’s quite difficult. But if you can engineer a subset of the data you can address the problem in a much more targeted way. When you talk about engineering the data, what do you mean exactly? Ng: In AI, data cleaning is important, but the way the data has been cleaned has often been in very manual ways. In computer vision, someone may visualize images through a Jupyter notebook and maybe spot the problem, and maybe fix it. But I’m excited about tools that allow you to have a very large data set, tools that draw your attention quickly and efficiently to the subset of data where, say, the labels are noisy. Or to quickly bring your attention to the one class among 100 classes where it would benefit you to collect more data. Collecting more data often helps, but if you try to collect more data for everything, that can be a very expensive activity. For example, I once figured out that a speech-recognition system was performing poorly when there was car noise in the background. Knowing that allowed me to collect more data with car noise in the background, rather than trying to collect more data for everything, which would have been expensive and slow. Back to top What about using synthetic data, is that often a good solution? Ng: I think synthetic data is an important tool in the tool chest of data-centric AI. At the NeurIPS workshop, Anima Anandkumar gave a great talk that touched on synthetic data. I think there are important uses of synthetic data that go beyond just being a preprocessing step for increasing the data set for a learning algorithm. I’d love to see more tools to let developers use synthetic data generation as part of the closed loop of iterative machine learning development. Do you mean that synthetic data would allow you to try the model on more data sets? Ng: Not really. Here’s an example. Let’s say you’re trying to detect defects in a smartphone casing. There are many different types of defects on smartphones. It could be a scratch, a dent, pit marks, discoloration of the material, other types of blemishes. If you train the model and then find through error analysis that it’s doing well overall but it’s performing poorly on pit marks, then synthetic data generation allows you to address the problem in a more targeted way. You could generate more data just for the pit-mark category. “In the consumer software Internet, we could train a handful of machine-learning models to serve a billion users. In manufacturing, you might have 10,000 manufacturers building 10,000 custom AI models.” —Andrew Ng Synthetic data generation is a very powerful tool, but there are many simpler tools that I will often try first. Such as data augmentation, improving labeling consistency, or just asking a factory to collect more data. Back to top To make these issues more concrete, can you walk me through an example? When a company approaches Landing AI and says it has a problem with visual inspection, how do you onboard them and work toward deployment? Ng: When a customer approaches us we usually have a conversation about their inspection problem and look at a few images to verify that the problem is feasible with computer vision. Assuming it is, we ask them to upload the data to the LandingLens platform. We often advise them on the methodology of data-centric AI and help them label the data. One of the foci of Landing AI is to empower manufacturing companies to do the machine learning work themselves. A lot of our work is making sure the software is fast and easy to use. Through the iterative process of machine learning development, we advise customers on things like how to train models on the platform, when and how to improve the labeling of data so the performance of the model improves. Our training and software supports them all the way through deploying the trained model to an edge device in the factory. How do you deal with changing needs? If products change or lighting conditions change in the factory, can the model keep up? Ng: It varies by manufacturer. There is data drift in many contexts. But there are some manufacturers that have been running the same manufacturing line for 20 years now with few changes, so they don’t expect changes in the next five years. Those stable environments make things easier. For other manufacturers, we provide tools to flag when there’s a significant data-drift issue. I find it really important to empower manufacturing customers to correct data, retrain, and update the model. Because if something changes and it’s 3 a.m. in the United States, I want them to be able to adapt their learning algorithm right away to maintain operations. In the consumer software Internet, we could train a handful of machine-learning models to serve a billion users. In manufacturing, you might have 10,000 manufacturers building 10,000 custom AI models. The challenge is, how do you do that without Landing AI having to hire 10,000 machine learning specialists? So you’re saying that to make it scale, you have to empower customers to do a lot of the training and other work. Ng: Yes, exactly! This is an industry-wide problem in AI, not just in manufacturing. Look at health care. Every hospital has its own slightly different format for electronic health records. How can every hospital train its own custom AI model? Expecting every hospital’s IT personnel to invent new neural-network architectures is unrealistic. The only way out of this dilemma is to build tools that empower the customers to build their own models by giving them tools to engineer the data and express their domain knowledge. That’s what Landing AI is executing in computer vision, and the field of AI needs other teams to execute this in other domains. Is there anything else you think it’s important for people to understand about the work you’re doing or the data-centric AI movement? Ng: In the last decade, the biggest shift in AI was a shift to deep learning. I think it’s quite possible that in this decade the biggest shift will be to data-centric AI. With the maturity of today’s neural network architectures, I think for a lot of the practical applications the bottleneck will be whether we can efficiently get the data we need to develop systems that work well. The data-centric AI movement has tremendous energy and momentum across the whole community. I hope more researchers and developers will jump in and work on it. Back to top This article appears in the April 2022 print issue as “Andrew Ng, AI Minimalist.”

뉴스

타임라인 키워드

Leftist climate zealotry means never having to say you’re sorry

‘Red Rocks’ Review: Weirdo, Cliff-Jumping Kiddies Are the Focus of Bruno Dumont’s Latest Experiment

CBS Isn’t Suppressing Stephen Colbert’s ‘Only in Monroe’ Public Access Show, It’s On His New YouTube Channel

Should wedding guests tip the bartender? Here’s what etiquette experts say you should do

‘No defense, only offense’: America’s growing antisemitism problem requires more than silence

Colbert Wasn't Funny

Lawsuit Against Virginia Tech Alleging Anti-Male Bias in Title IX Proceedings Can Go Forward in Part

Developers: Get Your Medical Mobile App Verified By IEEE

How Melbourne’s AI and Data Center Flywheel Is Accelerating Research Innovation

Accelerating Chipmaking Innovation for the Energy-Efficient AI Era

IEEE Program Aims to Connect the Billions Who Are Still Offline

Neutralizing the Gigascale Problem: How to Solve the Physical Power Paradox of Extreme AI Training Loads

With $1 Cyberattacks on the Rise, Durable Defenses Pay Off

Better Hardware Could Turn Zeros into AI Heroes

Andrew Ng: Unbiggen AI