A naukri.com initiative
AR News
Semiengineering
1M
432
Image Credit: Semiengineering
Benchmark and Evaluation Framework For Characterizing LLM Performance In Formal Verification (UC Berkeley, Nvidia)
- A technical paper titled 'FVEval: Understanding Language Model Capabilities in Formal Verification of Digital Hardware' has been published by researchers at UC Berkeley and NVIDIA.
- The paper introduces FVEval, a benchmark and evaluation framework for characterizing the performance of large language models (LLMs) in formal verification (FV) tasks related to digital chip design.
- FVEval consists of three sub-tasks that measure LLM capabilities at different levels, including generating SystemVerilog assertions, reasoning about design RTL, and suggesting assertions without human input.
- The paper evaluates a wide range of existing LLMs against FVEval, providing insights into the current state of LLM performance and its potential application in improving digital FV productivity.
Read Full Article
26 Likes
Skarredghost
1M
45
Image Credit: Skarredghost
Snap Spectacles 5 hands-on: a nice devkit towards a brighter future
- Snap Spectacles 5 is the latest iteration of Snap Spectacles, featuring waveguides and LCoS miniature projectors. Spectacles started off as a device for taking photos and small videos for Snapchat and have evolved to become an AR device that creators can use to build AR lenses for Snapchat.
- Spectacles 5 are currently being marketed as a dev kit for creators to rent for $99 per month, with the minimum commitment of 12 months. Specs include a 45min battery life, built-in stereo speakers, a six-microphone array for audio input, 2x full-color high-rez cameras, and 6-axis IMUs for inertial sensing.
- The device is focused on Spectacles app and Snapchat ecosystem, though some people may find other ways to monetize it indirectly, such as receiving funds from Snap or using it to build their personal brand.
- Visuals are affected by the same problems that all other AR glasses are having, including ‘rainbows’ due to waveguides and persistence. As for the quality of visuals, it is nice, but there are many visual artifacts due to the waveguides
- The hands-on testing notes that the device is bulky and heavy, and even though Snap has done much as it could, it is not possible to make the lightweight AR glasses of our dreams yet.
- Content is scant at present, but Snap hopes to build a vibrant platform if the commercial version of Spectacles is released in the future as a general-purpose device and not just a Snapchat add-on.
- The reviewer suggests that Snap should detach Spectacles from Snapchat, provide more choices for developers, and make sure that developers have a clear monetization path to create a rich ecosystem.
- Overall, Snap Spectacles 5 is a nice gadget that is projected to be a devkit towards a brighter future, being one step towards a consumer edition of Spectacles that may come in a few years, fixing all its current problems.
Read Full Article
2 Likes
Siliconangle
1M
22
Image Credit: Siliconangle
On theCUBE Pod: A look at connected ecosystems and questions of an AI election
- Connected ecosystems and real-time environments are defining the infrastructure game in the current era of business models.
- AI is creating a disruption while changing multiple things in the business model creating new questions around the business model itself.
- Companies are starting to move from classic software-as-a-service ecosystems on cloud to connected ecosystems.
- Real-time processing and data exchange creates a requirement for CapEx build-out and usage-based pricing models.
- SuperComputing, KubeCon and CloudNativeCon, are all big shows coming up that highlight the ongoing importance of infrastructure to the industry.
- Investment in AI is continuing, particularly from Amazon, which is reorganizing its startup group in anticipation of further growth.
- The 2024 election is shaping up to be defined by AI, with the question on everyone's mind being which side will be more effective with this emerging technology.
- From theCUBE, John Furrier and Dave Vellante believe that connected ecosystems and real-time processing are just the beginning of large and rapid changes to come.
- Join our community on YouTube for more tech news and updates.
- Support our mission to provide free, deep, and relevant content with just one click below.
Read Full Article
1 Like
Hackaday
1M
18
Image Credit: Hackaday
Pi Zero to AR: Building DIY Augmented Reality Glasses
- Redditor mi_kotalik has created 'Zero', DIY augmented reality (AR) glasses using a Raspberry Pi Zero.
- Zero offers features like video playback, Bluetooth audio, teleprompter, and image viewer, making it an affordable, self-contained AR device.
- The creation of Zero involved designing the frame, experimenting with lenses, and customizing SPI displays for real-time response.
- mi_kotalik plans to develop a V2 of Zero with a Compute Module 4 for enhanced capabilities like 3D rendering, GPS, and spatial tracking.
Read Full Article
1 Like
Immersivelearning
1M
451
How VR and AR are transforming education
- VR and AR technologies are transforming education by enhancing learning environments and making education more accessible, engaging, and effective.
- Benefits of VR and AR in education include enhanced engagement, accessibility to complex topics, and safe learning environments for students.
- Real-world case studies include Google Expeditions, which allows virtual field trips, and University of Illinois, which uses VR for medical training.
- The future of education with VR and AR includes personalized learning experiences, global collaboration, and lifelong learning.
Read Full Article
27 Likes
Insider
1M
150
Image Credit: Insider
Chip giant Nvidia wants to bring robots to the hospital
- Nvidia is developing physical AI to make robots for use in hospitals for applications such as X-rays and linen delivery.
- Nvidia has been investing in and partnering with digital health and biotech companies, aiming to bring the next wave of artificial intelligence to healthcare.
- Kimberly Powell, Nvidia's VP of Healthcare, said that Nvidia's investments aim to corner the emerging advanced robotics market in order to maintain its spot in tech.
- Nvidia's future 'physical AI' will enable machines to perceive, understand and interact with the physical world.
- The development of physical AI will allow hospital systems to train their clinicians on virtual models by creating a digital twin; this could help them design healthcare systems.
- Nvidia is partnering with IT solutions provider Mark III to create digital simulations of hospital environments to tackle the digital twin challenge in healthcare.
- Nvidia has invested in Moon Surgical, a robotics company, to provide an extra set of arms during surgeries.
- Mayo Clinic has been deploying its own linen delivery robots, which ferry fresh sheets from room to room.
- Despite investing heavily in digital health companies, Nvidia does not plan on acquiring any healthcare firms.
- Nvidia is teaming up with Microsoft to bring AI tools and expert support to start-ups in healthcare and life sciences.
Read Full Article
9 Likes
Pymnts
1M
432
Image Credit: Pymnts
Why Companies Want Accounts Receivables to Get Smarter and Faster Right Now
- Many B2B businesses are facing challenges with their manual accounts receivable (AR) processes, leading to delayed payments, errors, and increased operational costs.
- Automating AR workflows through technology can improve financial resilience, operational efficiency, and cash flow, while reducing manual tasks and allowing finance teams to focus on strategic initiatives.
- Despite the benefits, only about 24% of companies have implemented dedicated AR automation software, with many still relying on outdated manual processes.
- AR automation, coupled with artificial intelligence (AI) and machine learning (ML) technologies, can provide actionable insights, predict cash flow bottlenecks, and optimize customer interactions.
Read Full Article
26 Likes
Cgmagonline
1M
396
Image Credit: Cgmagonline
NVIDIA GeForce NOW: Games To Be Added In November 2024
- NVIDIA GeForce NOW will receive 17 new titles in November 2024.
- Dragon Age: The Veilguard and Resident Evil 4 (2023) are among the games being added.
- Dragon Age: The Veilguard is a new release by Electronic Arts and BioWare.
- Resident Evil 4 has been remade with updated graphics and gameplay.
Read Full Article
23 Likes
Eu-Startups
1M
127
Berlin-based Plato secures €6 million aiming to transform the wholesale industry with AI
- Plato, an AI-powered ERP automation platform for wholesale distributors, has raised €6 million in pre-seed funding.
- The funding round was led by Cherry Ventures, with support from the German government and tech industry leaders.
- Plato's platform integrates with existing ERP systems, using AI to enhance sales and operational workflows.
- The funding will be used to expand Plato's team and scale its product across multiple large customers.
Read Full Article
7 Likes
Pymnts
1M
1.1k
Image Credit: Pymnts
Nvidia Reportedly Weighing Investment in Musk’s xAI
- Nvidia is reportedly considering investing in Elon Musk's AI startup, xAI.
- xAI is aiming to raise billions of dollars, valuing the company at around $40 billion.
- Musk has been in discussions with tech companies and venture firms for funding.
- xAI is expecting a major funding round in January, potentially valuing the company at $75 billion.
Read Full Article
23 Likes
Gizchina
1M
401
Image Credit: Gizchina
Rumored High-end Nvidia CPU Will Reportedly Launch Next Year
- Nvidia is rumored to launch its first high-end CPU in September 2025, marking its entry into the consumer CPU market.
- The CPU is rumored to be Arm-based and will combine Nvidia's GPU technology with an innovative CPU platform.
- The initial release will target high-end PCs, with a broader rollout expected in March 2026.
- Nvidia's entry into the CPU market could reshape the competitive landscape currently dominated by Intel and AMD.
Read Full Article
24 Likes
Medium
1M
387
Image Credit: Medium
Lost in Time: Where Would You Travel?
- If given the opportunity to travel in time, the author would choose to go to the future.
- The author is specifically interested in the advancements in technology, such as augmented reality and computer implants.
- The author envisions a future where augmented reality is seamlessly integrated into glasses or contact lenses and computer implants are commonly used.
- In this future, the author imagines a sporting event that combines physical reality with augmented images, creating a unique and interactive experience.
Read Full Article
23 Likes
Dev
1M
337
Image Credit: Dev
How to run for inference Llama-3_1-Nemotron-51B-Instruct?
- The LLM model, Llama-3_1-Nemotron-51B-Instruct, was developed by NVIDIA using Neural Architecture Search (NAS) to balance model efficiency and correctness.
- The NAS technique eliminated superfluous elements to produce a more efficient architecture for effective inference on the H100 GPU.
- Knowledge distillation was also applied in creating the Nemotron-51B student model from the larger Llama-3.1-70B teacher model, preserving accuracy and significantly reducing the model's size.
- The Nemotron model was shown to provide an excellent accuracy-efficiency tradeoff with lower computational costs and excellent performance.
- For deploying the Llama-3_1-Nemotron-51B-Instruct model, GPU-powered Virtual Machine offered by NodeShift is used.
- In this tutorial, the reader was taken through a step-by-step guide to deploying Llama-3_1-Nemotron-51B-Instruct on a GPU-powered virtual machine with NodeShift.
- Prerequisites include GPUs such as A100 80GB or H100, at least 100GB RAM, and 150GB free disk space.
- The tutorial covers account setup, creation of a GPU Node, model selection, region selection, storage selection, authentication method selection, image selection, package and library installation, model loading, and generating responses.
- NodeShift provides an accessible, secure, and affordable platform to run AI models efficiently, and is an excellent choice for users looking to experiment with Llama-3_1-Nemotron-51B-Instruct and other cutting-edge AI tools.
- The tutorial concludes by providing links to NodeShift resources such as their website, docs, LinkedIn, Discord, and daily.dev.
Read Full Article
20 Likes
Siliconangle
1M
351
Image Credit: Siliconangle
Why Jamie Dimon is Sam Altman’s biggest competitor
- Investors have poured north of $30 billion into independent foundation model players focusing on the false grail of so-called artificial general intelligence (AGI). However, we believe the greatest value capture exists in what is called Enterprise AGI. Private enterprises will capture the majority of value in the race for AI leadership. This report will dig deep into the economics of foundation models and will analyze Enterprise AGI and the untapped opportunities that exist within enterprises.
- The false grail of AGI is often associated with OpenAI Chief Executive Sam Altman, driven by the pursuit of machine business work better than humans. However, removing humans' involvement for decision-making is not as easy as it seems.
- AGI in the enterprise is the ability gradually to learn and adapt white-collar work processes of the firm. Instead of one all-intelligent AGI, it is a swarm of modestly intelligent agents that can collectively augment human white-collar work. Nvidia CEO’s quotes also resonate with this idea of a directory of AIs, some digital, some biological, specialized and skilled, and just generally good at doing things. An all-knowing AGI may not be a viable scenario.
- Private enterprises possess unique data and process advantage that is not in the public domain, which is the key ingredient for its competitive advantage. The proprietary knowledge specific to a business resides in the enterprise's data estate, with foundation models deriving from the transformer, differentiated solely by the data sets that train it.
- LLMs are focused on scaling up its algorithm without recognizing that scaling up Models does not mean scaling their accuracy. The cost of scaling models is becoming unsustainable, and prices will drop like a rock. The investments going to foundation models are misguided, and the massive value capture exists in Enterprise AGI, which is a large opportunity for companies and most of the value created will accrue to these firms.
- Citizen developers give agents goals and guardrails. But just as important, agents can generate plans with step-by-step reasoning that humans can then edit and iterate. Then the exception conditions become learnable moments to help get the agent further down the long tail the next time. Agents can learn from their human supervisors and learn from exceptions while in production. This is the opposite of traditional software where you want to catch and suppress bugs before production. This extension of what an agent can do can create a potentially learnable step to extend the long tail of activities and handle more edge cases.
- Agents can change the economics of achieving that by observing and learning from human actions. The swarm of workflow agents, which are really specialized action models, collectively have the ability to learn and embody all that management knowhow and outperform the most advanced foundation model. The collective intelligence of those agents outperforms the singular intelligence in a frontier foundation model.
- The enterprise automation opportunity is significant but requires specialized knowledge because each process is unique, and it takes more than an AI model to understand an enterprise's processes. The future of enterprise software must not only deliver transactional efficiency and productivity but also be capable of managing and analyzing the complexities of real-time business processes in a unified framework subject to operational, analytic and historical systems.
- Though there are various emerging players exploring the integrated, process-driven source of truth, understanding the nuances of each approach will be critical to building applications that can capitalize on this integrated, process-driven source of truth. The future presents both challenges and a significant opportunity, setting the stage for a new era in enterprise applications, data orchestration, and process alignment.
- Overall, we believe that Enterprise AGI provides a massive opportunity for companies, and most of the value created will accrue to these firms. Private enterprises will capture the majority of value in the race for AI leadership, underscoring the importance of Enterprises' ability to gradually learn and adapt white-collar work processes over an all-knowing AGI.
Read Full Article
21 Likes
TechCrunch
1M
173
Image Credit: TechCrunch
Quantum Machines and Nvidia use machine learning to get closer to an error-corrected quantum computer
- Quantum Machines and Nvidia are getting one step closer to an error-corrected quantum computer by using machine learning on Nvidia’s DGX Quantum computing platform to keep the system calibrated. The π pulses that control the rotation of a qubit inside a quantum processor are being calibrated using an off-the-shelf reinforcement learning model. This control problem lends itself to being solved with the help of reinforcement learning because a quantum system is always slightly different. A small improvement in calibration can lead to massive improvements in error correction of a logical qubit.
- The actual code for running the experiment was only about 150 lines long. All of the work the two teams did to integrate the various systems and build out the software stack, though, can be hidden away from developers. Szmuk stressed that for this project, the team only worked with a very basic quantum circuit but that it can be generalized to deep circuits as well.
- Quantum error correction is a huge problem that is necessary to unlock fault-tolerant quantum computing. DGX Quantum is the first system that enables the kind of minimal latency needed to perform these calculations. As quantum computers scale up, problems become bottlenecks that are really compute-intensive. Applying exactly the right control pulses to get the most out of the qubits is one of the problems.
- The team is only at the start of this optimization process and collaboration and expects to create more and more open-source libraries over time to take advantage of this larger platform. With Nvidia’s Blackwell chips becoming available next year, they’ll also have an even more powerful computing platform for this project.
- If you look at the performance of quantum computers today, you get some high fidelity. Then, the users, when they use the computer, it’s typically not at the best fidelity. It drifts all the time. If we can frequently recalibrate it using these kinds of techniques and underlying hardware, then we can improve the performance and keep the fidelity high over a long time.
- Quantum Machine’s co-founder and CTO, Yonatan Cohen, noted how his company has long sought to use general classical compute engines to control quantum processors, but those compute engines were small and limited. The partnership with Nvidia and their DGX platform brings that computational power to calibration.
- The collaboration between Quantum Machines and Nvidia is a small step towards solving the most important problems. Useful quantum computing is going to require the tight integration of accelerated supercomputing, which may be the most difficult engineering challenge.
- Sam Stanwyck, Nvidia’s group product manager for quantum computing, stated that the two companies plan to continue this collaboration and get these tools into the hands of more researchers.
- The team used TD3 algorithm for calibration, because it worked best. The team only worked with a very basic quantum circuit, but it can be generalized to deep circuits as well.
- Constantly adjusting π pulses in near real-time is an extremely compute-intensive task requiring strong computational power, but collaborative team working between Quantum Machines and Nvidia is making it possible.
Read Full Article
10 Likes
For uninterrupted reading, download the app