Milton Luaces

AI 🤖: Muy importante! 👇 Adopting and implementing GenAI successfully starts with your people. It requires reimagining your workplace culture at every level of the organisation to one that supports innovation and trust. Nice one Bryan McGowan! #ResposibleAI #TrustedAI

3

SAS Viya integrates seamlessly with existing systems and data infrastructures so you avoid the costly and disruptive process of “rip and replace” (completely removing and replacing existing systems). #AI #Analytics #Innovation #SystemRedesign

3

The following article is curated and summarized by ChatGPT: Google DeepMind's NATURAL PLAN is a benchmark designed to assess large language models

1

Organizations across the world have invested in and continue to invest in data modernization as a foundational core pillar of their digital transformation. More lately, data has been elevated from an asset status to being served as a product that aligns with business value streams. However, data leaders across industries acknowledge one gap : data value quantification. Specifically when it comes to quantifying the value of these investments to business, data leaders find themselves without a framework for value quantification. Any modern data strategy formulation needs to have a component or approach that helps to quantify the value that businesses will derive through the execution of this strategy. #ChiefDataOfficer #CDO #DataStrategy #ITStrategy #CIO

12

SEAL is a new model benchmarking service from Scale AI that fills a big gap: an additional objective, third-party, trustworthy way to compare models. People distrust many benchmarks because they worry that models are optimized to perform well on them—including by training on the underlying data (like cheating on a test by memorizing the questions and answers in advance). The LMSys Chatbot Arena (https://chat.lmsys.org) addresses this by having humans evaluate model output in blinded head-to-head tests. The new SEAL leaderboard (https://lnkd.in/gdCZzRnz) does something similar, but uses "thoroughly vetted experts using domain specific methodologies" to do the head-to-head evaluation. Sounds much more expensive, but they just raised $1 billion, so I think they can afford it.

1

💡 Transforming Training Intake With AI: An ROI-Driven Approach💡 I'm thrilled to present on this topic with my (much smarter and better-looking) colleague Scott Provence at #ATDCore4 in Phoenix this September. Will you join us at the conference to boost your skills in design, delivery, evaluation, and learning tech? Secure your spot today and enjoy early bird savings before August 2! #ATDCore4

17

2 commentaires

Hi Streamlit-lovers! 😍 Happy to 𝐨𝐩𝐞𝐧-𝐬𝐨𝐮𝐫𝐜𝐞 a side-project I developed to improve Streamlit chat UI! 💬 (already leveraged at Roquette) 👇 - 📥 𝐂𝐡𝐚𝐭 𝐡𝐢𝐬𝐭𝐨𝐫𝐲 𝐞𝐱𝐩𝐨𝐫𝐭 (in JSON format) - 🧹 𝐂𝐥𝐞𝐚𝐫 𝐜𝐡𝐚𝐭 - 🔁 𝐑𝐞𝐫𝐮𝐧 𝐥𝐚𝐬𝐭 𝐪𝐮𝐞𝐬𝐭𝐢𝐨𝐧 - 👍 and 👎 𝐟𝐞𝐞𝐝𝐛𝐚𝐜𝐤 𝐛𝐮𝐭𝐭𝐨𝐧𝐬 (registered in the logs) - 💬 𝐭𝐨𝐤𝐞𝐧𝐬 𝐜𝐨𝐮𝐧𝐭 - 👩‍💻 𝐜𝐡𝐚𝐭 𝐚𝐯𝐚𝐭𝐚𝐫𝐬 𝐩𝐞𝐫𝐬𝐨𝐧𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 📚 𝐆𝐢𝐭𝐇𝐮𝐛 𝐑𝐞𝐩𝐨: https://lnkd.in/er9PjYFc

38

4 commentaires

The future of AI education should include hands-on experience with private AI agents. With tools like Ollama and phidata setting up your own AI agent requires minimal coding - think under 10 lines. Perfect for beginners and enthusiasts. Integrating such practical knowledge into standard education could transform how we understand and interact with AI technology. Github: https://lnkd.in/g2fpAQDP #AI #Education #Technology #Innovation

5

Building robust #data capabilities is essential for moving from reactive to proactive #sustainability. At #CDOIQ2024, Christine Legner from the University of Lausanne shared key findings on how companies can better manage sustainability-related data. From sourcing practices to #datastewardship, Christine provided a framework that helps organizations address the challenges of sustainability reporting and the requirements of the EU’s Corporate Sustainability Reporting Directive (#CSRD). These insights are crucial for any organization aiming to achieve long-term sustainability goals. More about the session ➡️https://lnkd.in/gsAXwZ93 #CDO #CDOIQ #datamanagement #AI #datagovernance #datascience #analytics

9

Do business decision makers trust your ML models ? 'While machine learning models offer significant advantages, their increasing complexity poses challenges regarding explainability, interpretability, and compliance, impacting their adoption and effectiveness. Techniques, such as SHAP and feature importance empower data scientists to understand their models better, therefore accommodating the process of explaining the predictions to the business' #explainableAI #AI #ceo #cfo #cmo # https://lnkd.in/e6uYve3s

6

1 commentaire

#MeasureSummit is underway again… Our Measurement and Data team are looking forward in particular to the sessions on: - AI-nalytics: Practical Uses of AI in Analytics. - Beyond the Basics: Solving Digital Analytics Challenges with AI and Advanced Modelling. Sessions also covering POAS based tracking for Google, under the bonnet of consent mode v2, marketing attribution, quantifying messaging and measuring attitudes, server-side tracking and CRO reports in GA4. Expecting another interesting, challenging and thought-provoking couple of days.

11

1 commentaire

Just published my latest article: "Transforming Business Operations with LLMs: A Path to a Production Ecosystem." a breakdown of the key components necessary for integrating LLMs into core business operations, including Human In The Loop (HITL), Feedback, Evaluation, and Decision Executor subsystems. Learn how these elements work together to start building an AI platform.

37

13 commentaires

Collibra's Pushdown for Snowflake is transforming how organisations manage data quality by enabling efficient, scalable processing directly within the cloud. This integration not only enhances data governance but also optimises performance and resource use, ensuring your data remains accurate and reliable. Learn more about how this powerful combination can improve your data strategy on our blog: https://hubs.la/Q02L-KNs0 #DataQuality #Collibra #Snowflake #DataGovernance

10

From Data to Decisions: Key Skills for Modern Analytics In a world increasingly generating data, the importance of hybrid skills within Data, Analytics, and AI functions cannot be overstated. Our roles are fundamentally about creating value by distilling vast amounts of data into consumable, actionable insights. To achieve this, we must excel across four key pillars: business acumen, data storytelling, data analytics skills, and technical expertise. Business Acumen: * Data & Analytics Alignment: Partnering with business leaders to understand and demonstrate how analytics can be leveraged to solve business problems. * Macro-Perspective: Understanding the company’s business strategy, current business issues, priorities, and industry trends. * Business Knowledge: Grasping business frameworks, strategies, operations, processes, and data decision points. Storytelling: * Data Storytelling: Crafting compelling narratives with data. Keep it simple and easy to consume. * Insight Articulation: Explaining data-driven insights to highlight current and predicted recommendations, their impact, and opportunities. * Soft Skills: Utilising communication and interpersonal skills to effectively convey insights gained from analysis. Data Analytics Skills: * Data Analysis: Evaluating data using analytical and logical reasoning to discover insights. * Data Capabilities: Understanding the theory and application of data , analytics, and enablement tools. * Mathematical: Understanding significance of outputs. Technical Skills: * Data Modelling: Structuring data to enable the analysis of information, both internal and external to the business. * Querying: Querying and manipulating data to solve complex problems. * Testing & Validating: Defining, developing, and implementing quality assurance practices and procedures for technical solutions and validating hypotheses. By mastering these hybrid skills, professionals in Data, Analytics, and AI can effectively bridge the gap between technical expertise and business strategy, driving meaningful insights and fostering innovation. #empowerpeople #AI #analytics

72

2 commentaires

Been conflicted for a while about the utility of #Ai in the data space. #Ai rocks compared to #humans for repetitive task automation, pattern recognition and anomaly detection, analytics, multi-variable optimization problems, writing papers and creating art. But we don't see self-driving cars everywhere; there's a reason a driver must be present ... the general public, policy-makers and insurance companies feel more at ease with the #aihuman teamup. Check out the movie by Netflix called "Atlas" starring Jennifer Lopez https://lnkd.in/gMDdJ8p4 ("Atlas Shepherd (Lopez), a brilliant but misanthropic data analyst with a deep distrust of artificial intelligence") Is Ai reliable enough for cleaning #mrodata used in critical business operations? Ai is #aidelusional and #aischizophrenic when the #aiprompts or #aitraining data are "bad" or when the #algorithm isn't well-tuned. The #investment and #roi of Ai and the #losttime can cause #buyersremorse. Worse, especially if #baddata causes a #blackswan event. In that case, who do we blame, Ai or the #executiveteam? #Ai alone cannot solve #everything. If you test it, you'll find mistakes that then must be corrected by humans - this is the #aigap - the investment in Ai is not small, and because Ai makes mistakes, the output needs to be reviewed. Mistakes identified then must be fixed by humans, and the data Ai cannot handle must also be processed by humans ... we can use a process that is time-tested and known to work very well --> https://lnkd.in/emyCncUq (Sphera's Data Quality Optimization Service). Regarding highly critical #MRO #masterdata I'm not ready to fully #trust Ai, because there's too much at stake. There's far too much #risk. For non-critical, simple data Ai is fine - some mistakes are OK; plants will not blow up, supply chain will function (if not without errors), inventory may not be perfect, and sustainability might suffer a little. For everything else (that is #critical), make sure humans are involved with a bullet-proof process for curation, creation and quality control. Employ time-tested #datastandards #datagovernance #datacleansing #mdm and #dataquality. This way, improved outcomes in #supplychain, #inventory, #plantmaintenance, #processsafety and #sustainability can happen. At the end of the day, it's about five (5) things: (1) Quality, (2) Schedule, (3) Experience, (4) Service, and (5) Price. If #baddata is hurting your business, it's time to get it right. Check out @Sphera https://lnkd.in/gEyWmwGB or ping Dave Kuketz for a consultation.

16

Don't miss this exclusive interview with the Godfather of Data Strategy! (yes, really).   Step into the captivating world of data leadership with Peter Jackson , the esteemed Chief Data & Technology Officer at Outra and published authority on Data Strategy.   Discover a wealth of insights from this seasoned industry veteran as he unveils transformative data strategies across diverse sectors.   Read more: https://lnkd.in/g36WX9nP #DataLeadership #DataStrategy #Innovation #DataGovernance #DataAnalytics #DataInsights #DataQuality #DataArchitecture #DigitalInnovation #ShitStorm

42

6 commentaires