ML6

🌟 Join us at Beam Summit 2024, an in-person global conference in sunny California, featuring keynotes, talks, and networking focused on the latest trends in .

We’re excited to contribute with two insightful sessions:

◾September 4, 15:00 | multi-modal LLM data processing with Apache Beam: https://hubs.la/Q02MRFpq0
◾September 5, 13:30 | RAG data ingestion using Apache Beam: https://hubs.la/Q02MRG0K0

Don’t miss this chance to connect, share ideas, and be part of the vibrant data community!

Join us for an exclusive workshop on AI Governance by ML6 and at the AI innovation Center, High Tech Campus on 24 September.🧭

Our experts will cover:
✔️AI Governance from Theory to Practice
✔️The EU AI Act: what it means for you
✔️Practical tips for implementation

Don't miss out! Register now: https://hubs.la/Q02M0fTP0

The final text of the European AI Act was published two weeks ago, introducing new AI system regulations. The first step to compliance? List and classify your AI systems. 🗂

We created a decision-tree to help, which we are already using with clients.

Want to try it? Download it here: https://hubs.la/Q02LR_V60

Our legal counsel Michiel Van Lerbeirghe was featured in De Standaard this week, shedding light on the critical intersection of AI, privacy, and intellectual property law in Europe. 🔍

As Michiel puts it: "The big question is how to apply existing laws, such as the legislation of intellectual property rights and the GDPR, to these new developments." The outcome of ongoing legal cases and disputes, such as Schrems' (and his privacy organisation's) recent complaint against X, will be interesting for defining the boundaries of AI development and usage.

Read the full article here [dutch]: https://hubs.la/Q02L6V0R0

Klacht tegen X omdat het chatbot Grok traint met Europese posts De Oostenrijkse privacy-activist Max Schrems dient klacht in bij de Belgische gegevensbeschermingsautoriteit omdat X zonder toestemming de berichten van Europeanen gebruikt om de AI-chatbot Grok te trainen.

🍦It's getting hot out there! As we keep on building a diverse team at ML6, we are always looking for new flavours! Which flavour are you? You can even be the sprinkles on top!

Flashback to our fantastic workshop with ! 💥 We traded theory for hands-on action, turning attendees into guardrail gurus with NVIDIA's NeMo. With this initiative we want to foster and encourage diversity in technology and help accelerate talent in AI further.

Check out this video on implementing NVIDIA NeMo Guardrails for essential safety measures and prompt injections: https://hubs.ly/Q02Jx7Kz0

Ready to make your LLMs safer and smarter? Stay tuned for our post-summer meetup schedule!

Since the launch of ChatGPT, Generative AI tools have been widely adopted in organisations, yet they bring concerns such as deep fakes, ethical dilemmas, copyright issues, and privacy risks. Organisations are responding by implementing AI governance practices, including policies, guidelines and risk management frameworks, to identify and mitigate these risks effectively. In this blog post, we delve into existing risk management frameworks and we share our learnings based on our hands-on experience.

https://hubs.la/Q02HWj2D0

🚀 Microsoft Copilot: Discover RAG Made Easy(?)!

In recent years, Retrieval Augmented Generation RAG) solutions have enhanced the way we can interact with Large Language Models, enhancing their abilities across a number of use-cases. From Microsoft we have Copilot, positioning itself as the everyman's solution to what can be a complicated process.

💡 What's Inside?

◾Explore how RAG (Retrieval-Augmented Generation) are enhancing natural language generation.
◾Learn about Microsoft's Copilot, a versatile helper chatbot, and its integration within the Microsoft 365 ecosystem.
◾Understand the various tools like Copilot Studio, Power Automate, Bot SDK Framework, and Azure AI Studio to customize and extend your AI capabilities.

📈 Why Read This?
Gain insights into the latest AI trends, practical applications, and how these technologies can streamline your operations. Whether you're looking for quick integration or deep customization, this post has you covered.

🔗 You can read the full article here: https://hubs.la/Q02HBQBp0

📢 Join the conversation, share your thoughts, and stay ahead in the AI revolution!

🚨Join us for our webinar 'AI Governance insights: all you need to know about the AI Act’ !🚨

With the upcoming AI regulations and rapidly evolving tech landscape, we are excited to announce the launch of our new webinar series 'AI Governance Insights'. In this series we will discuss how AI systems are built, deployed, used and managed in a responsible and compliant way.

⚖️ We open the series with our first webinar, 'AI Governance insights: all you need to know about the AI Act'. We'll cover 10 key topics, including how rules vary by your organisation's role, classification of prohibited and high-risk AI systems, and transparency obligations for AI models.

Don't miss out! Join us on the 27th of August at 12:00 CEST. You can register here: https://hubs.la/Q02Hcx3n0

We are proud of our diverse workforce, with 14 nationalities represented at ML6. 🤩😎

🚀 Unlocking the Future of Protein Design with Machine Learning: A Deep Dive into Sequence Generation.

Are you passionate about the fascinating world of protein design? Then our latest blog post is a must-read! At ML6, we’ve delved into the cutting-edge advancements in machine learning that are revolutionising the field, specifically focusing on protein sequence generation. 🌟

🔍 Why Proteins Matter:
Proteins are the building blocks of life, playing crucial roles in virtually every cellular process. From providing structural support to facilitating enzymatic activity, their importance cannot be overstated. Our comprehensive overview explains why mastering protein manipulation could unlock solutions to some of the world's most pressing environmental challenges.

💡 The ML Revolution:
Machine learning has been intertwined with protein research since the 1980s, and recent technological advancements have supercharged these efforts. We explore how modern ML approaches, including Diffusion Models and Large Language Models (LLMs), are being harnessed to predict and generate protein sequences with unprecedented accuracy.

🌐 What’s New:
Protein Sequence Generation: Discover how LLMs and diffusion models are transforming the creation of new proteins.
Innovative Models: Get insights into the latest models pushing the boundaries of what’s possible in protein design.
Practical Takeaways: Learn about the real-world applications and potential of these groundbreaking techniques.

📢 Dive into our blog to explore how these advancements are not just theoretical but are paving the way for real-world applications that could change the game in biotechnology and beyond. Read now and stay ahead in the ever-evolving intersection of protein design and machine learning! Your expertise could be the next big leap forward.

👉 https://hubs.la/Q02GV9L50

Ever heard of a chatbot selling a car for $1? 🤖🚗 Believe it or not, someone managed to convince a customer service chatbot to sell a car at this price through clever prompting. Such incidents are part of a growing wave of attacks as LLMs become integrated into diverse applications. Dive into the world of LLM Guardrails with us and learn how to prevent such incidents in your applications!

📅 Date: 24th July 2024
📍 Location: Online + ML6 Amsterdam office (hybrid)

Join us and PyLadies Amsterdam for an interactive session where we’ll dive into the world of LLM Guardrails. Discover the mechanisms that ensure applications produce reliable, robust, safe, and ethical outputs, and understand their crucial role in LLMs. We’ll cover everything from content moderation to prompt injection and other essential safety measures.

But it’s not just theory – you’ll get hands-on experience implementing your own guardrails using tools like NVIDIA’s NeMo Guardrails. By the end of the workshop, you’ll be a guardrail guru, ready to make your LLM applications safer, more accurate, and robust.

👩‍💻 This event celebrates diversity in tech – welcoming participants of all Python proficiency levels without regard to labels (gender, sexuality, or race). Whether you’re an engineer, data scientist, or AI enthusiast, join us to innovate responsibly and inclusively! 🚀

With this initiative, ML6 loves to support diversity in technology by collaborating with PyLadies Amsterdam. Secure your spot now and expand your network in the tech community. Don’t miss out! https://hubs.la/Q02GHL2w0

The EU AI Act is finally published! ✔️ What does it mean?

◾August 2024: The Act will enter into force.
◾February 2025: Obligations for prohibited AI systems (❗Facial & emotions recognition in some cases)
◾August 2025: GPAI model rules start
◾August 2026: Obligations for high risk AI systems apply (❗Employment, education and government in some cases)
◾August 2027: Obligations for AI systems in products apply"

Source: https://hubs.la/Q02GlH8M0

🔬 Revolutionizing Protein Design with ESM3: A Leap Forward in Biotechnology

Are you involved in drug discovery, enzyme engineering, or biotechnology? Understanding the latest advancements in protein design is crucial for staying ahead in these fields. Dive into our latest blog post on ESM3 - The Frontier of Protein Design and discover how this groundbreaking model is transforming the landscape of protein science.

🌟 Why It Matters:

ESM3 unifies sequence, structure, and function prediction, surpassing previous models that could only specialize in one domain.
Learn how ESM3’s multi-input/multi-output architecture sets a new standard for protein generation, accelerating the discovery of novel proteins with tailored functionalities.

💡 Key Takeaways:

Bidirectional Transformer Architecture: Captures long-range dependencies crucial for understanding protein structure and function.
Geometric Attention: Enhances spatial awareness, leading to more accurate protein structure predictions.
Chain of Thought Generation: Ensures comprehensive understanding by alternating between sequence and structure during generation.

🔍 Performance Highlights:

Generates high-confidence structures (pTM > 0.8, pLDDT > 0.8) with low sequence similarity to known proteins, pushing generative possibilities into unknown territory.
Showcases an outstanding ability to plausibly generate 'distant' proteins, opening new avenues for innovative biomolecular designs.

🔗 Discover More:

Explore how ESM3 can revolutionize your approach to protein design and bring innovative solutions to complex biomedical challenges.
Read the full blog post here: https://hubs.la/Q02FyjRq0

🤝 Engage with Us:

Share your thoughts, leave comments, and let's discuss how ESM3 can impact your work. Don’t forget to share this post with your network to spread the knowledge!

Stay ahead in the field of protein science and biotechnology with ESM3. 🌍🔬

Responsible AI should be an integral part of every AI solution. At ML6, we want to develop solutions that are legally compliant, ethically responsible, and technically secure. In this blog post, we transparently share how we have integrated Responsible AI into ML6's processes.

https://hubs.la/Q02Ff6L10

From LLM judges to chain-of-thought: How can we ensure that LLM-based applications are safe, reliable, and ethically responsible?

In our latest blogpost, we give high-level overview of guardrail techniques, their perks and pitfalls. Language Models (LLMs) are increasingly used for applications, but also come with risks like leaking confidential data and spreading misinformation. Especially now that LLM-based applications are more and more deployed for public use, implementing guardrails becomes crucial to ensure the safety and reliability. 🔐

Read the full blogpost: https://hubs.la/Q02D_WpN0

Meet ESM3, the state-of-the-art language model transforming how we understand and engineer proteins.
Curious about how AI can achieve unprecedented accuracy in predicting and designing proteins? ESM3 leverages advanced AI techniques like bi-directional transformers and geometric attention to unify protein sequence generation, structure prediction, and function prediction.

The impact is significant: ESM3's precision in predicting protein structures and functions is revolutionizing biomedical research. It's accelerating drug development, enabling precise protein engineering, and uncovering new biotechnological applications. 🧪

A highlight of ESM3's capabilities is its ability to generate novel proteins, such as esmGFP, a fluorescent protein unlike any found in nature. This showcases ESM3's innovative power in designing functional biomolecules.
Join us in exploring the future of protein modeling with ESM3! Stay tuned for our upcoming blog post detailing how this technology is reshaping the biotech landscape.

https://hubs.ly/Q02DpV0Z0

EvolutionaryScale Launches with ESM3: A Milestone AI Model for Biology [MEDIA ALERT]ESM3 Simulates 500 Million Years of Evolution to Create New Fluorescent Protein

🚀 As we keep on growing and creating business impact, we keep on welcoming new talent to accelerate intelligence together! Meet Iris, Sofie, Nicholas, Zakaria, William, Kais and Axel. Welcome aboard!

We're celebrating a major milestone - hitting 15,000 followers on LinkedIn! 🎉

A huge thank you to our incredible community for your continued support and engagement. We love sharing our expertise and learning from all of you. Here's to many more connections, conversations, and growing together!

Ghent is named Belgium's AI Capital! 🚀

De Tijd has recognized Ghent as Belgium's AI hub! Having been part of this thriving tech community for over10 years, we've witnessed its remarkable growth. We're thrilled to see more innovative companies joining the scene, including our own spin-offs like Paperbox and uman!

🚀 Case Study: https://hubs.la/Q02z2Njl0 & ML6 - Revolutionizing digital journalism

📈 Impact
ntv Nachrichten an RTL brand and top German news portal, partners with ML6 to optimize its editorial process using machine learning. This collaboration enhances workflows, improves content quality, and boosts daily news production, leading the way in modern journalism.

🚧 Challenges
Facing industry disruptions and evolving consumption habits, ntv Nachrichten aims to maintain its status and grow advertising revenue by enhancing digital operations and output quality.

💡 Solution
Teaming with ML6, ntv Nachrichten integrates Large Language Models (LLMs) to streamline editorial processes. This system efficiently assesses and generates news articles, with editors maintaining final control to ensure quality.

🏆 Results
The ML6-powered solution reduces production time, increases daily news output, and enhances content quality. This efficiency allows journalists more time for creative tasks, optimizing the editorial team's resources.

Discover the full case study on how ntv Nachrichten is shaping the future of journalism on our website: https://hubs.la/Q02z2NVX0

🎮 Who are the trophy hunters among us? We've been busy lately and always on the look for talent to join and create real business impact with AI! Going for platinum?

If you’re immersed in the world of AI, the recent developments around the European AI Act are likely on your radar. If not, it’s the right time to proactively think about the impact of the AI Act on your (future) AI project. While many provisions won't be enforced immediately, gaining an early understanding of the AI Act will not only help reduce the risk of future fines but also facilitate the development of responsible AI solutions and prevent the need for project modifications once the AI Act is fully enforced. 🇪🇺

In this article, we explore how to navigate the requirements for developing high-risk AI systems within the regulation of the European AI Act, ensuring compliance and fostering ethical AI innovation. 💡

https://hubs.la/Q02xcQym0

💊Can Existing Drugs Fight HIV?

HIV remains a complex and challenging disease, despite extensive research. Traditional drug development is slow and costly. This blog post explores drug repurposing: using existing, safe drugs for new applications. 🔎

Our intern, Andrew, leveraged data from HIV and utilized 3 innovative approaches to identify potential drugs. These drugs, originally designed for other purposes, may hold promise in the fight against HIV.

Intrigued? 👉 Learn more about our approach and the potential of drug repurposing in HIV treatment: https://hubs.la/Q02vQC4B0