Jin Huang

NetAI is at the forefront of AIOps innovation, utilizing Graph Neural Networks (GNNs) to revolutionize the industry. By harnessing the power of GNNs, NetAI can efficiently and accurately analyze complex interconnected data sources, extracting valuable insights. This innovative approach enables NetAI to map and model intricate relationships within network data, empowering organizations to proactively address issues before they escalate. The result is enhanced operational efficiency and a higher level of accuracy in both detecting and resolving problems. With NetAI paving the way in AIOps advancement, organizations can adopt a smarter and more proactive approach to managing their network infrastructure. Contact me to arrange a detailed demonstration mike@netai.ai

26

Navigating the Complexity of Similarity Measurements and Latent Space Abstraction in Large Text Corpora In the realm of large text corpora, both similarity measurements and model abstraction of latent spaces present significant challenges. These tasks, while distinct, share an underlying complexity that demands a deep understanding of the nuances in natural language processing (NLP). In fact, similarity measurement is almost as tricky as developing a large language model (LLM) itself. Similarity measurements in large text corpora involve determining how closely related different pieces of text are. On the surface, this might seem straightforward—count the common words or phrases, measure distances in a vector space, and you're done. However, the reality is far more complex. Textual similarity is influenced by syntax, semantics, context, and even the cultural or domain connotations of words. Traditional methods like cosine similarity or the Jaccard index often fall short when dealing with nuanced texts, as they may not fully capture the contextual relationships between words and phrases. This challenge is heightened when the text corpus is vast, diverse, and rich in idiomatic or domain-specific language. On the other hand, model abstraction of latent space refers to how a model, such as an LLM, organizes and represents information internally. Latent space is where the model abstracts complex data into vectors that encapsulate meaning, allowing it to understand and generate human-like text. The abstraction process is deeply intricate, as the model must balance between overfitting to specific details and generalizing across broader contexts. Creating effective latent spaces requires a fine-tuned architecture that can capture the subtleties of language, including metaphor, irony, implied meanings, and specific context. These spaces are multidimensional and not easily interpretable, making it challenging to ensure they represent text in a way that aligns with human intuition. The complexity here rivals, if not surpasses, the challenges faced in similarity measurements. In essence, similarity measurements are not just about counting overlaps or calculating distances; they are about understanding relationships within a rich and complex linguistic space. This requirement places them on par with the intricacies of developing and fine-tuning LLMs, where nuance in language can shift the balance between a model's success and failure.

5

1 Comment

Open Source AI should provide the general public with access to design, develop, and implement their AI solutions. This can be achieved by developing a programming paradigm that suits the perspective of every entity. At some point everyone needs to communicate with another entity. All entities other than ourselves are truly black-box, Human or AI. We as Humans do not have telepathy, so we communicate verbally and nonverbally in the physical world. This communication is not always effective with both Humans and AI, in both directions: transmitting and receiving. With AI and Humans we have a black box that can receive our transmitted input and maybe might decide to provide, or transmit back, a relevant output. Really nothing has changed. Human or AI we are both ultimately semi- non-deterministic black boxes. So, what does that mean? We focus our attention on providing the infrastructure for AI driven development. AI Programming Paradigm, AI Programming Language, AI Compiler, AI general processing units, and of course an AI operating system. That being said Open Source AI has a lot of work ahead of it. It is a good time to exist.

1

TRIPLE BAM! If you missed it, here's the recording of our live Q/A with Jay Alammar and Joshua Starmer PhD! We talked about many AI topics, answered lots of great questions, and talked about two great books: Hands-On Large Language Models: Language Understanding and Generation, and The Illustrated Guide to Neural Networks and AI!!! https://lnkd.in/dcPHs7YQ

44

1 Comment

AI has become sentient in Japan https://lnkd.in/gh9GKD22

1

🚀 Our new blog series on AI for #digitalthreads and #digitalengineering will share practical examples and use cases for leveraging #generativeAI and #llms for digital threads. Part 1 shows you how to connect with an AI service (OpenAI) in < 5 mins from the Syndeia Web Dashboard. Power of #APIs and #Syndeia's #nocode Generic RESTful Integration capability.

17

🚨 BREAKING NEWS! 🚨 🔥 Latttice + Claude.ai = A NEW ERA in Data Exploration! 🔥 🚀 Get ready to supercharge your data with the power of AI! With Claude.ai integrated into Latttice, accessing and exploring data has never been this EASY or EXCITING! 💥 💬 Talk to your data! Ask questions in plain English and get instant, actionable insights. No more complicated queries—just results! 🔗 Seamless integration with all your data, NO coding required. It’s as simple as having a conversation. 🤖 🔒 Security that’s enterprise-grade—because your data deserves the best protection. 🛡️ 💡 Say goodbye to guesswork and hello to data-driven brilliance! Ready to take your decision-making to the next level? 🌟 👉 Join our Beta Program TODAY and experience the future of data firsthand! #AIRevolution #DataExploration #ClaudeAI #Latttice #NoCode #SmarterDecisions #TechInnovation #FutureOfData #AI #DataScience #BetaTest

3

1 Comment

Just over a year in and GPT4 is now 12x cheaper, faster, better! We're most excited about GPT-4o – for complex reasoning tasks, and Gemini 1.5 Flash – for data transformations, especially given its long context length. As chips get faster, model architectures improve, and we have more data, it feel inevitable that the cost of intelligence (as a service) will keep going down -- just like what happened for compute and storage costs in the last decades.

77

2 Comments

Let’s be real. Building an AI agent isn’t the hard part anymore. The real challenge is making it perform well in real-world scenarios. Optimization, accuracy, and reliability are what truly matter. I am focusing on creating AI agents that do more than just function. They should be reliable, cost-efficient, and scalable. If you’re working on AI agents, remember that the details make all the difference. Focusing on these harder aspects is what will set your solution apart. What do you think? Let’s discuss in the comments. #AI #ArtificialIntelligence #AIAgents #Optimization #Halosphere #TechInnovation

6

Startups live or die by their ability to execute at speed. For large companies, too, the speed with which an innovation team is able to iterate has a huge impact on its odds of success. Generative AI makes it possible to quickly prototype AI capabilities. AI capabilities that used to take months can sometimes be built in days or hours by simply prompting a large language model. I find this speed exciting and have been thinking about how to help startups and large companies alike go faster. I’ve been obsessed with speedy execution for a long time. When working on a project, I am loath to take two weeks to do something that I could do in one week. The price of moving at that pace is not that we take one week longer (which might be okay) but that we’re 2x slower (which is not)! When building an AI-powered product, there are many steps in designing, building, shipping, and scaling the product that are distinct from building the AI capability, and our ability to execute these other steps has not sped up as much as the AI part. But the speed with which we can prototype AI creates significant pressure to speed up these other steps, too. If it took 6 months to collect data, train a supervised learning algorithm, and deploy the model to the cloud, it might be okay to take 2 months to get user feedback. But if it takes a week to build a prototype, waiting 2 months for feedback seems intolerably slow! I’d like to focus on one key step of building applications: getting user feedback. A core part of the iterative workflow of designing and building a product (popularized by Eric Ries in his book The Lean Startup) is to build a prototype (or MVP, minimum viable product), get user feedback on it, and to use that feedback to drive improvements. The faster you can move through this loop — which may require many iterations — the faster you can design a product that fits the market. This is why AI Fund, a venture studio that I lead, uses many fast, scrappy tactics to get feedback. For B2C (business to consumer) offerings, here is a menu of some options for getting customer feedback: 1. Ask 3 friends or team members to look at the product and let you know what they think (this might take ~0.5 days). 2. Ask 10 friends or team members to take a look (~2 days). 3. Send it to 100 trusted/volunteer alpha testers (~1 week?). 4. Send it to 1,000 users to get qualitative or quantitative feedback (~2 weeks?). 5. Incorporate it into an existing product to get feedback (1 to 2 months?). 6. Roll it out to a large user base of an existing product and do rigorous A/B testing. As we go down this list, we get (probably) more accurate feedback, but the time needed to get that feedback increases significantly. Also, the tactics at the top of the list create basically no risk, and thus it’s safe to repeatedly call on them, even with preliminary ideas and prototypes. [Reached length limit. Full text: https://lnkd.in/gEBBB2UZ ]

4,913

211 Comments

2024 has been a game-changer for spatial computing 🚀 With the Vision Pro dropping, Quest 3 making waves - we're seeing this tech leap from "maybe one day" to "happening right now." What's wild is how different industries are running with it: 1️⃣ Surgeons practicing procedures in 3D 2️⃣ Kids exploring the solar system from their classroom 3️⃣ Trying on clothes without stepping in a store And property? That's where we've been putting in the work. Ensuring everyone has the chance to experience their can step inside their future home before it's built. To Explore their space in lifelike detail across any device, from smartphone to vr headset. Started this journey years ago because I saw how painful it was trying to sell off plans and blueprints. Now? We're watching people walk through unbuilt spaces like they're already there. Across any device they've got. But here's what keeps me up at night (in a good way) - this feels like we're just scratching the surface. What space do you think spatial computing will shake up next? #SpatialComputing #PropTech #CGI #ArchViz

20

~$1.2M of grants now available for helping us build various aspects of neural-symbolic-evolutionary AGI on OpenCog Hyperon platform ... See https://lnkd.in/d_Gn9yRA .... Transformer NNs are super cool but only a piece of the AGI puzzle. We are recruiting AI devs to work on a number of the other pieces, which we have figured out how to piece toward AGI using the PRIMUS design on Hyperon platform, centered on our new MeTTa AGI language. Let's get rational, compassionate, beneficial, decentralized and democratized AGI built first -- before Big Tech and Big Gov't get too close to AGI with their centralized AI systems lacking any deep understanding of self and other and objectifying people toward narrow goals.

221

21 Comments

Kudos to Kshitiz Parashar for his pioneering work on Agentic Chunking! His latest primer has quickly become the must-read guide for AI Application Engineers looking to optimize LLMs for smarter chunking, better retrieval, and sharper accuracy. This Speaks volumes about the calibre of the team which I am super proud of.  Even a lean, dedicated group can now harness the power of LLMs to design state-of-the-art architectures—proof that AI is leveling the playing field like never before. Ready to dive in? Explore the full article here: https://lnkd.in/eJ8cem3x #aiagents #chunking

37

1 Comment

Meet AGUVIS: A pure vision-based framework for autonomous Computer/Mobile-use GUI agents, operating seamlessly across web, desktop, and mobile platforms without UI code. Key Features & Contributions - Pure Vision Framework: First fully autonomous pure vision GUI agent capable of performing tasks independently without relying on closed-source models - Cross-Platform Unification: Unified action space and plugin system that works consistently across different GUI environments - Comprehensive Dataset: Large-scale dataset of GUI agent trajectories with multimodal grounding and reasoning - Two-Stage Training: Novel training pipeline focusing on GUI grounding followed by planning and reasoning - Inner Monologue: Explicit planning and reasoning capabilities integrated into the model training - Generalization to OOD (Out-of-Distribution): AGUVIS demonstrates impressive generalization on OSWorld benchmark even without desktop trajectory data during training. Project Page: https://lnkd.in/guzkwpbB Paper: https://lnkd.in/gs7wXyJS GitHub: https://lnkd.in/gWeSkFmK Collaboratiion among Yiheng Xu Zekun Wang Junli Wang Dunjie Lu Tianbao Xie Amrita Saha Doyen Sahoo Tao Yu Caiming Xiong

155

6 Comments

Differentiable Adaptive Merging is accelerating SLMs for enterprises - https://buff.ly/4fiUK4A #AI #genAI #LLMs

1

Generative AI: miracle worker or sophisticated guesser? While AI can now write, code, and create, it's essentially still a prediction machine. This revelation challenges how we view and utilize AI in business. What does this mean for your AI strategy? Let's discuss. #ArtificialIntelligence #BusinessStrategy #TechInnovation #AILimitations #FutureOfWork

2