A common problem with brand tracking is that while online surveys are relatively cheap, the respondents are usually not representative of the general population but can be heavily biased. Additionally, clients are often interested in how their brand fares in specific target groups, for instance households with young children. It can be especially tricky to get reliable estimates for very small target groups.
Bayesian methods can help solve these problems but they are still not widely used in production: Bayesian models are commonly thought of as being too slow and too compute-intensive to use them at scale.
In this presentation, I talk about how a Bayesian approach addresses these problems and helps us to deliver more accurate results to our clients. By using AWS Batch together with the Netflix library Metaflow, we can run more than a thousand models in parallel per project thus allowing us to scale our operations.
Talk given at the Berlin Bayesians, 10/11/2021
Taking Machine Learning from Batch to Real-Time (big data eXposed 2015)Elad Rosenheim
This is my talk given at eXelate's "big data eXposed 2015" event in Tel-Aviv.
It touches the common problem of trying to optimize websites for better conversion rates, and covers the range of solutions - starting at classic A/B testing (unsuitable for rapidly changing content), going through Multi-Arm Bandits and finally arriving at Contextual Bandits - online machine learning algorithms allowing for true personalization in real-time.
Recommender Systems from A to Z – Model TrainingCrossing Minds
This second meetup will be about training different models for our recommender system. We will review the simple models we can build as a baseline. After that, we will present the recommender system as an optimization problem and discuss different training losses. We will mention linear models and matrix factorization techniques. We will end the presentation with a simple introduction to non-linear models and deep learning.
Système de recommandations de produits sur un site marchand par Koby KARP, Data Scientist (Equancy) & Hervé MIGNOT, Partner at Equancy
La recommandation reste un outil clé pour la personnalisation des sites marchands et le sujet est loin d’être épuisé. La prise en compte de la particularité d’un marché peut nécessité d’adapter le traitement et les algorithmes utilisés. Après une revue des techniques de recommandations, nous présenterons la démarche spécifique que nous avons adopté. Le système a été développé sous Spark pour la préparation des données et le calcul des modèles de recommandations. Une API simple et son service ont été développé pour délivrer les recommandations aux applications clientes.
Building machine learning service in your business — Eric Chen (Uber) @PAPIs ...PAPIs.io
When making machine learning applications in Uber, we identified a sequence of common practices and painful procedures, and thus built a machine learning platform as a service. We here present the key components to build such a scalable and reliable machine learning service which serves both our online and offline data processing needs.
Sangchul Song and Thu Kyaw discuss machine learning at AOL, and the challenges and solutions they encountered when trying to train a large number of machine learning models using Hadoop. Algorithms including SVM and packages like Mahout are discussed. Finally, they discuss their analytics pipeline, which includes some custom components used to interoperate with a range of machine learning libraries, as well as integration with the query language Pig.
This Machine Learning Algorithms presentation will help you learn you what machine learning is, and the various ways in which you can use machine learning to solve a problem. At the end, you will see a demo on linear regression, logistic regression, decision tree and random forest. This Machine Learning Algorithms presentation is designed for beginners to make them understand how to implement the different Machine Learning Algorithms.
Below topics are covered in this Machine Learning Algorithms Presentation:
1. Real world applications of Machine Learning
2. What is Machine Learning?
3. Processes involved in Machine Learning
4. Type of Machine Learning Algorithms
5. Popular Algorithms with a hands-on demo
- Linear regression
- Logistic regression
- Decision tree and Random forest
- N Nearest neighbor
What is Machine Learning: Machine Learning is an application of Artificial Intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.
- - - - - - - -
About Simplilearn Machine Learning course:
A form of artificial intelligence, Machine Learning is revolutionizing the world of computing as well as all people’s digital interactions. Machine Learning powers such innovative automated technologies as recommendation engines, facial recognition, fraud protection and even self-driving cars.This Machine Learning course prepares engineers, data scientists and other professionals with knowledge and hands-on skills required for certification and job competency in Machine Learning.
- - - - - - -
Why learn Machine Learning?
Machine Learning is taking over the world- and with that, there is a growing need among companies for professionals to know the ins and outs of Machine Learning
The Machine Learning market size is expected to grow from USD 1.03 Billion in 2016 to USD 8.81 Billion by 2022, at a Compound Annual Growth Rate (CAGR) of 44.1% during the forecast period.
- - - - - -
What skills will you learn from this Machine Learning course?
By the end of this Machine Learning course, you will be able to:
1. Master the concepts of supervised, unsupervised and reinforcement learning concepts and modeling.
2. Gain practical mastery over principles, algorithms, and applications of Machine Learning through a hands-on approach which includes working on 28 projects and one capstone project.
3. Acquire thorough knowledge of the mathematical and heuristic aspects of Machine Learning.
4. Understand the concepts and operation of support vector machines, kernel SVM, naive Bayes, decision tree classifier, random forest classifier, logistic regression, K-nearest neighbors, K-means clustering and more.
5. Be able to model a wide variety of robust Machine Learning algorithms including deep learning, clustering, and recommendation systems
- - - - - - -
How to design your ML application to be production ready from the day one
How to switch from notebooks to deployable and maintainable software
How to deploy, serve and monitor prediction pipelines
How to re-train models in production
How to shift machine learning experimentation phase to production
Taking Machine Learning from Batch to Real-Time (big data eXposed 2015)Elad Rosenheim
This is my talk given at eXelate's "big data eXposed 2015" event in Tel-Aviv.
It touches the common problem of trying to optimize websites for better conversion rates, and covers the range of solutions - starting at classic A/B testing (unsuitable for rapidly changing content), going through Multi-Arm Bandits and finally arriving at Contextual Bandits - online machine learning algorithms allowing for true personalization in real-time.
Recommender Systems from A to Z – Model TrainingCrossing Minds
This second meetup will be about training different models for our recommender system. We will review the simple models we can build as a baseline. After that, we will present the recommender system as an optimization problem and discuss different training losses. We will mention linear models and matrix factorization techniques. We will end the presentation with a simple introduction to non-linear models and deep learning.
Système de recommandations de produits sur un site marchand par Koby KARP, Data Scientist (Equancy) & Hervé MIGNOT, Partner at Equancy
La recommandation reste un outil clé pour la personnalisation des sites marchands et le sujet est loin d’être épuisé. La prise en compte de la particularité d’un marché peut nécessité d’adapter le traitement et les algorithmes utilisés. Après une revue des techniques de recommandations, nous présenterons la démarche spécifique que nous avons adopté. Le système a été développé sous Spark pour la préparation des données et le calcul des modèles de recommandations. Une API simple et son service ont été développé pour délivrer les recommandations aux applications clientes.
Building machine learning service in your business — Eric Chen (Uber) @PAPIs ...PAPIs.io
When making machine learning applications in Uber, we identified a sequence of common practices and painful procedures, and thus built a machine learning platform as a service. We here present the key components to build such a scalable and reliable machine learning service which serves both our online and offline data processing needs.
Sangchul Song and Thu Kyaw discuss machine learning at AOL, and the challenges and solutions they encountered when trying to train a large number of machine learning models using Hadoop. Algorithms including SVM and packages like Mahout are discussed. Finally, they discuss their analytics pipeline, which includes some custom components used to interoperate with a range of machine learning libraries, as well as integration with the query language Pig.
This Machine Learning Algorithms presentation will help you learn you what machine learning is, and the various ways in which you can use machine learning to solve a problem. At the end, you will see a demo on linear regression, logistic regression, decision tree and random forest. This Machine Learning Algorithms presentation is designed for beginners to make them understand how to implement the different Machine Learning Algorithms.
Below topics are covered in this Machine Learning Algorithms Presentation:
1. Real world applications of Machine Learning
2. What is Machine Learning?
3. Processes involved in Machine Learning
4. Type of Machine Learning Algorithms
5. Popular Algorithms with a hands-on demo
- Linear regression
- Logistic regression
- Decision tree and Random forest
- N Nearest neighbor
What is Machine Learning: Machine Learning is an application of Artificial Intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.
- - - - - - - -
About Simplilearn Machine Learning course:
A form of artificial intelligence, Machine Learning is revolutionizing the world of computing as well as all people’s digital interactions. Machine Learning powers such innovative automated technologies as recommendation engines, facial recognition, fraud protection and even self-driving cars.This Machine Learning course prepares engineers, data scientists and other professionals with knowledge and hands-on skills required for certification and job competency in Machine Learning.
- - - - - - -
Why learn Machine Learning?
Machine Learning is taking over the world- and with that, there is a growing need among companies for professionals to know the ins and outs of Machine Learning
The Machine Learning market size is expected to grow from USD 1.03 Billion in 2016 to USD 8.81 Billion by 2022, at a Compound Annual Growth Rate (CAGR) of 44.1% during the forecast period.
- - - - - -
What skills will you learn from this Machine Learning course?
By the end of this Machine Learning course, you will be able to:
1. Master the concepts of supervised, unsupervised and reinforcement learning concepts and modeling.
2. Gain practical mastery over principles, algorithms, and applications of Machine Learning through a hands-on approach which includes working on 28 projects and one capstone project.
3. Acquire thorough knowledge of the mathematical and heuristic aspects of Machine Learning.
4. Understand the concepts and operation of support vector machines, kernel SVM, naive Bayes, decision tree classifier, random forest classifier, logistic regression, K-nearest neighbors, K-means clustering and more.
5. Be able to model a wide variety of robust Machine Learning algorithms including deep learning, clustering, and recommendation systems
- - - - - - -
How to design your ML application to be production ready from the day one
How to switch from notebooks to deployable and maintainable software
How to deploy, serve and monitor prediction pipelines
How to re-train models in production
How to shift machine learning experimentation phase to production
Understanding computer vision with Deep LearningCloudxLab
Computer vision is a branch of computer science which deals with recognising objects, people and identifying patterns in visuals. It is basically analogous to the vision of an animal.
Topics covered:
1. Overview of Machine Learning
2. Basics of Deep Learning
3. What is computer vision and its use-cases?
4. Various algorithms used in Computer Vision (mostly CNN)
5. Live hands-on demo of either Auto Cameraman or Face recognition system
6. What next?
Understanding computer vision with Deep Learningknowbigdata
Computer vision is a branch of computer science which deals with recognising objects, people and identifying patterns in visuals. It is basically analogous to the vision of an animal.
Topics covered:
1. Overview of Machine Learning
2. Basics of Deep Learning
3. What is computer vision and its use-cases?
4. Various algorithms used in Computer Vision (mostly CNN)
5. Live hands-on demo of either Auto Cameraman or Face recognition system
6. What next?
Understanding computer vision with Deep LearningShubhWadekar
Topics covered in the Webinar
1. Overview of Machine Learning
2. Basics of Deep Learning
3. What is computer vision and its use-cases?
4. Various algorithms used in Computer Vision (mostly CNN)
5. Live hands-on demo of either Auto Cameraman or Face recognition system
6. What next?
Presented by Sandeep Giri
www.cloudxlab.com
[PythonPH] Transforming the call center with Text mining and Deep learning (C...Paul Lo
Transforming the call center with Text mining and Deep learning:
1. Text ming tool to unlock user insights
2. Artificial Intelligence revolution in call centers: deep learning-based bot
Disclaimer: This talk is about the models, processes, and algorithms that power machine learning APIs rather than about the API itself.
Machine learning is fast becoming ubiquitous, from large enterprises to small startups. In this talk, we’ll embark on a journey with a small risk management startup, and we’ll talk about how snowflakes & trees and chickens & eggs came together to help the startup implement its own flavor of machine learning. This is the story of the cold decision tree and how it solved the cold start problem.
Managing the Machine Learning Lifecycle with MLflowDatabricks
ML development brings many new complexities beyond the traditional software development lifecycle. MLflow is an open-source project from Databricks aiming to solve some of these challenges such as experiment tracking, reproducibility, model packaging, deployment, and governance, in order to manage and accelerate the lifecycle of your ML projects.
Mariia Havrylovych "Active learning and weak supervision in NLP projects"Fwdays
Successful artificial intelligence solutions always require a massive amount of high-quality labeled data. In most cases, we don’t have a large and qualitative labeled set together. Weak supervision and active learning tools may help you optimize the labeling process and address the shortage of data labels.
First, we will review how active learning can significantly reduce the amount of labeled data for training with classic approaches. We will show how active learning methods can be customized for a specific (NLP) task by using text embedding.
With weak supervision, we will see how using simple rules gets a big train dataset automatically and high model performance without manual labeling at all.
In the end, we will combine active learning and weak supervision by taking advantage of both techniques and achieving the best metrics.
Strategic AI Integration in Engineering TeamsUXDXConf
This presentation dives into the practical applications of machine learning within Google's operations, providing a comprehensive overview of how to leverage AI technologies to solve real-world business challenges.
Key Points Covered:
- Introduction to Machine Learning at Google: Discussion on the role of ML and its evolution in enhancing Google's operational efficiency.
- Experience Sharing: Insights into the team's long-term engagement with machine learning projects and the impacts on Google’s operational strategies.
- Practical Applications: Real-world examples of ML applications within Google’s daily operations, providing a blueprint to adapt similar strategies.
- Challenges and Solutions: Discussion on the challenges faced during the implementation of ML projects and the strategic solutions employed to overcome them.
- Future of ML at Google: Insights into future trends in machine learning at Google and how they plan to continue integrating AI into their ecosystem.
In this talk we explore how to build Machine Learning Systems that can that can learn "continuously" from their mistakes (feedback loop) and adapt to an evolving data distribution.
The youtube link to video of the talk is here:
https://www.youtube.com/watch?v=VtBvmrmMJaI
Balancing Automation and Explanation in Machine LearningDatabricks
For a machine learning application to be successful, it is not enough to give highly accurate predictions: Customers also want to know why the model has made that prediction, so they can compare it against their intuition and (hopefully) gain trust in the model. However, there is a trade-off between model accuracy and explainability - for example, the more complex your feature transformations become, the harder it is to explain what the resulting features mean to the end customer. However, with the right system design this doesn't mean it has to be a binary choice between these two goals. It is possible to combine complex, even automatic, feature engineering with highly accurate models and explanations. We will describe how we are using lineage tracing to solve this issue at Salesforce Einstein, allowing good model explanations to coexist with automatic feature engineering and model selection. By building this into an open source AutoML library TransmogrifAI, an extension to SparkMlLib, it is easy to ensure a consistent level of transparency in all of our ML applications. As model explanations are provided out of the box, data scientists don't need to re-invent the wheel when model explanations need to be surfaced.
Automatic Forecasting using Prophet, Databricks, Delta Lake and MLflowDatabricks
As Atlassian continues to scale to more and more customers, the demand for our legendary support continues to grow. Atlassian needs to maintain balance between the staffing levels needed to service this increasing support ticket volume with the budgetary constraints needed to keep the business healthy – automated ticket volume forecasting is at the centre of this delicate balance
SMM Cheap - No. 1 SMM panel in the worldsmmpanel567
Boost your social media marketing with our SMM Panel services offering SMM Cheap services! Get cost-effective services for your business and increase followers, likes, and engagement across all social media platforms. Get affordable services perfect for businesses and influencers looking to increase their social proof. See how cheap SMM strategies can help improve your social media presence and be a pro at the social media game.
Everyone knows the power of stories, but when asked to come up with them, we struggle. Either we second guess ourselves as to the story's relevance, or we just come up blank and can't think of any. Unlocking Everyday Narratives: The Power of Storytelling in Marketing will teach you how to recognize stories in the moment and to recall forgotten moments that your audience needs to hear.
Key Takeaways:
Understand Why Personal Stories Connect Better
How To Remember Forgotten Stories
How To Use Customer Experiences As Stories For Your Brand
More Related Content
Similar to Brand tracking with Bayesian Models and Metaflow
Understanding computer vision with Deep LearningCloudxLab
Computer vision is a branch of computer science which deals with recognising objects, people and identifying patterns in visuals. It is basically analogous to the vision of an animal.
Topics covered:
1. Overview of Machine Learning
2. Basics of Deep Learning
3. What is computer vision and its use-cases?
4. Various algorithms used in Computer Vision (mostly CNN)
5. Live hands-on demo of either Auto Cameraman or Face recognition system
6. What next?
Understanding computer vision with Deep Learningknowbigdata
Computer vision is a branch of computer science which deals with recognising objects, people and identifying patterns in visuals. It is basically analogous to the vision of an animal.
Topics covered:
1. Overview of Machine Learning
2. Basics of Deep Learning
3. What is computer vision and its use-cases?
4. Various algorithms used in Computer Vision (mostly CNN)
5. Live hands-on demo of either Auto Cameraman or Face recognition system
6. What next?
Understanding computer vision with Deep LearningShubhWadekar
Topics covered in the Webinar
1. Overview of Machine Learning
2. Basics of Deep Learning
3. What is computer vision and its use-cases?
4. Various algorithms used in Computer Vision (mostly CNN)
5. Live hands-on demo of either Auto Cameraman or Face recognition system
6. What next?
Presented by Sandeep Giri
www.cloudxlab.com
[PythonPH] Transforming the call center with Text mining and Deep learning (C...Paul Lo
Transforming the call center with Text mining and Deep learning:
1. Text ming tool to unlock user insights
2. Artificial Intelligence revolution in call centers: deep learning-based bot
Disclaimer: This talk is about the models, processes, and algorithms that power machine learning APIs rather than about the API itself.
Machine learning is fast becoming ubiquitous, from large enterprises to small startups. In this talk, we’ll embark on a journey with a small risk management startup, and we’ll talk about how snowflakes & trees and chickens & eggs came together to help the startup implement its own flavor of machine learning. This is the story of the cold decision tree and how it solved the cold start problem.
Managing the Machine Learning Lifecycle with MLflowDatabricks
ML development brings many new complexities beyond the traditional software development lifecycle. MLflow is an open-source project from Databricks aiming to solve some of these challenges such as experiment tracking, reproducibility, model packaging, deployment, and governance, in order to manage and accelerate the lifecycle of your ML projects.
Mariia Havrylovych "Active learning and weak supervision in NLP projects"Fwdays
Successful artificial intelligence solutions always require a massive amount of high-quality labeled data. In most cases, we don’t have a large and qualitative labeled set together. Weak supervision and active learning tools may help you optimize the labeling process and address the shortage of data labels.
First, we will review how active learning can significantly reduce the amount of labeled data for training with classic approaches. We will show how active learning methods can be customized for a specific (NLP) task by using text embedding.
With weak supervision, we will see how using simple rules gets a big train dataset automatically and high model performance without manual labeling at all.
In the end, we will combine active learning and weak supervision by taking advantage of both techniques and achieving the best metrics.
Strategic AI Integration in Engineering TeamsUXDXConf
This presentation dives into the practical applications of machine learning within Google's operations, providing a comprehensive overview of how to leverage AI technologies to solve real-world business challenges.
Key Points Covered:
- Introduction to Machine Learning at Google: Discussion on the role of ML and its evolution in enhancing Google's operational efficiency.
- Experience Sharing: Insights into the team's long-term engagement with machine learning projects and the impacts on Google’s operational strategies.
- Practical Applications: Real-world examples of ML applications within Google’s daily operations, providing a blueprint to adapt similar strategies.
- Challenges and Solutions: Discussion on the challenges faced during the implementation of ML projects and the strategic solutions employed to overcome them.
- Future of ML at Google: Insights into future trends in machine learning at Google and how they plan to continue integrating AI into their ecosystem.
In this talk we explore how to build Machine Learning Systems that can that can learn "continuously" from their mistakes (feedback loop) and adapt to an evolving data distribution.
The youtube link to video of the talk is here:
https://www.youtube.com/watch?v=VtBvmrmMJaI
Balancing Automation and Explanation in Machine LearningDatabricks
For a machine learning application to be successful, it is not enough to give highly accurate predictions: Customers also want to know why the model has made that prediction, so they can compare it against their intuition and (hopefully) gain trust in the model. However, there is a trade-off between model accuracy and explainability - for example, the more complex your feature transformations become, the harder it is to explain what the resulting features mean to the end customer. However, with the right system design this doesn't mean it has to be a binary choice between these two goals. It is possible to combine complex, even automatic, feature engineering with highly accurate models and explanations. We will describe how we are using lineage tracing to solve this issue at Salesforce Einstein, allowing good model explanations to coexist with automatic feature engineering and model selection. By building this into an open source AutoML library TransmogrifAI, an extension to SparkMlLib, it is easy to ensure a consistent level of transparency in all of our ML applications. As model explanations are provided out of the box, data scientists don't need to re-invent the wheel when model explanations need to be surfaced.
Automatic Forecasting using Prophet, Databricks, Delta Lake and MLflowDatabricks
As Atlassian continues to scale to more and more customers, the demand for our legendary support continues to grow. Atlassian needs to maintain balance between the staffing levels needed to service this increasing support ticket volume with the budgetary constraints needed to keep the business healthy – automated ticket volume forecasting is at the centre of this delicate balance
Similar to Brand tracking with Bayesian Models and Metaflow (20)
SMM Cheap - No. 1 SMM panel in the worldsmmpanel567
Boost your social media marketing with our SMM Panel services offering SMM Cheap services! Get cost-effective services for your business and increase followers, likes, and engagement across all social media platforms. Get affordable services perfect for businesses and influencers looking to increase their social proof. See how cheap SMM strategies can help improve your social media presence and be a pro at the social media game.
Everyone knows the power of stories, but when asked to come up with them, we struggle. Either we second guess ourselves as to the story's relevance, or we just come up blank and can't think of any. Unlocking Everyday Narratives: The Power of Storytelling in Marketing will teach you how to recognize stories in the moment and to recall forgotten moments that your audience needs to hear.
Key Takeaways:
Understand Why Personal Stories Connect Better
How To Remember Forgotten Stories
How To Use Customer Experiences As Stories For Your Brand
Mastering Local SEO for Service Businesses in the AI Era is tailored specifically for local service providers like plumbers, dentists, and others seeking to dominate their local search landscape. This session delves into leveraging AI advancements to enhance your online visibility and search rankings through the Content Factory model, designed for creating high-impact, SEO-driven content. Discover the Dollar-a-Day advertising strategy, a cost-effective approach to boost your local SEO efforts and attract more customers with minimal investment. Gain practical insights on optimizing your online presence to meet the specific needs of local service seekers, ensuring your business not only appears but stands out in local searches. This concise, action-oriented workshop is your roadmap to navigating the complexities of digital marketing in the AI age, driving more leads, conversions, and ultimately, success for your local service business.
Key Takeaways:
Embrace AI for Local SEO: Learn to harness the power of AI technologies to optimize your website and content for local search. Understand the pivotal role AI plays in analyzing search trends and consumer behavior, enabling you to tailor your SEO strategies to meet the specific demands of your target local audience. Leverage the Content Factory Model: Discover the step-by-step process of creating SEO-optimized content at scale. This approach ensures a steady stream of high-quality content that engages local customers and boosts your search rankings. Get an action guide on implementing this model, complete with templates and scheduling strategies to maintain a consistent online presence. Maximize ROI with Dollar-a-Day Advertising: Dive into the cost-effective Dollar-a-Day advertising strategy that amplifies your visibility in local searches without breaking the bank. Learn how to strategically allocate your budget across platforms to target potential local customers effectively. The session includes an action guide on setting up, monitoring, and optimizing your ad campaigns to ensure maximum impact with minimal investment.
Top 3 Ways to Align Sales and Marketing Teams for Rapid GrowthDemandbase
In this session, Demandbase’s Stephanie Quinn, Sr. Director of Integrated and Digital Marketing, Devin Rosenberg, Director of Sales, and Kevin Rooney, Senior Director of Sales Development will share how sales and marketing shapes their day-to-day and what key areas are needed for true alignment.
Monthly Social Media News Update May 2024Andy Lambert
TL;DR. These are the three themes that stood out to us over the course of last month.
1️⃣ Social media is becoming increasingly significant for brand discovery. Marketers are now understanding the impact of social and budgets are shifting accordingly.
2️⃣ Instagram’s new algorithm and latest guidance will help us maintain organic growth. Instagram continues to evolve, but Reels remains the most crucial tool for growth.
3️⃣ Collaboration will help us unlock growth. Who we work with will define how fast we grow. Meta continues to evolve their Creator Marketplace and now TikTok are beginning to push ‘collabs’ more too.
Financial curveballs sent many American families reeling in 2023. Household budgets were squeezed by rising interest rates, surging prices on everyday goods, and a stagnating housing market. Consumers were feeling strapped. That sentiment, however, appears to be waning. The question is, to what extent?
To take the pulse of consumers’ feelings about their financial well-being ahead of a highly anticipated election, ThinkNow conducted a nationally representative quantitative survey. The survey highlights consumers’ hopes and anxieties as we move into 2024. Let's unpack the key findings to gain insights about where we stand.
Most small businesses struggle to see marketing results. In this session, we will eliminate any confusion about what to do next, solving your marketing problems so your business can thrive. You’ll learn how to create a foundational marketing OS (operating system) based on neuroscience and backed by real-world results. You’ll be taught how to develop deep customer connections, and how to have your CRM dynamically segment and sell at any stage in the customer’s journey. By the end of the session, you’ll remove confusion and chaos and replace it with clarity and confidence for long-term marketing success.
Key Takeaways:
• Uncover the power of a foundational marketing system that dynamically communicates with prospects and customers on autopilot.
• Harness neuroscience and Tribal Alignment to transform your communication strategies, turning potential clients into fans and those fans into loyal customers.
• Discover the art of automated segmentation, pinpointing your most lucrative customers and identifying the optimal moments for successful conversions.
• Streamline your business with a content production plan that eliminates guesswork, wasted time, and money.
The digital marketing industry is changing faster than ever and those who don’t adapt with the times are losing market share. Where should marketers be focusing their efforts? What strategies are the experts seeing get the best results? Get up-to-speed with the latest industry insights, trends and predictions for the future in this panel discussion with some leading digital marketing experts.
Is AI-Generated Content the Future of Content Creation?Cut-the-SaaS
Discover the transformative power of AI in content creation with our presentation, "Is AI-Generated Content the Future of Content Creation?" by Puran Parsani, CEO & Editor of Cut-The-SaaS. Learn how AI-generated content is revolutionizing marketing, publishing, education, healthcare, and finance by offering unprecedented efficiency, creativity, and scalability.
Understanding
AI-Generated Content:
AI-generated content includes text, images, videos, and audio produced by AI without direct human involvement. This technology leverages large datasets to create contextually relevant and coherent material, streamlining content production.
Key Benefits:
Content Creation: Rapidly generate high-quality content for blogs, articles, and social media.
Brainstorming: AI simulates conversations to inspire creative ideas.
Research Assistance: Efficiently summarize and research information.
Market Insights:
The content marketing industry is projected to grow to $17.6 billion by 2032, with AI-generated content expected to dominate over 55% of the market.
Case Study: CNET’s AI Content Controversy:
CNET’s use of AI for news articles led to public scrutiny due to factual inaccuracies, highlighting the need for transparency and human oversight.
Benefits Across Industries:
Marketing: Personalize content at scale and optimize engagement with predictive analytics.
Publishing: Automate content creation for faster publication cycles.
Education: Efficiently generate educational materials.
Healthcare: Create accurate content for patients and professionals.
Finance: Produce timely financial content for decision-making.
Challenges and Ethical Considerations:
Transparency: Disclose AI use to maintain trust.
Bias: Address potential AI biases with diverse datasets.
SEO: Ensure AI content meets SEO standards.
Quality: Maintain high standards to prevent misinformation.
Conclusion:
AI-generated content offers significant benefits in efficiency, personalization, and scalability. However, ethical considerations and quality assurance are crucial for responsible use. Explore the future of content creation with us and see how AI is transforming various industries.
Connect with Us:
Follow Cut-The-SaaS on LinkedIn, Instagram, YouTube, Twitter, and Medium. Visit cut-the-saas.com for more insights and resources.
It's another new era of digital and marketers are faced with making big bets on their digital strategy. If you are looking at modernizing your tech stack to support your digital evolution, there are a few can't miss (often overlooked) areas that should be part of every conversation. We'll cover setting your vision, avoiding siloes, adding a democratized approach to data strategy, localization, creating critical governance requirements and more. Attendees will walk away with actions they can take into initiatives they are running today and consider for the future.
How to Run Landing Page Tests On and Off Paid Social PlatformsVWO
Join us for an exclusive webinar featuring Mariate, Alexandra and Nima where we will unveil a comprehensive blueprint for crafting a successful paid media strategy focused on landing page testing.With escalating costs in paid advertising, understanding how to maximize each visitor’s experience is crucial for retention and conversion.
This session will dive into the methodologies for executing and analyzing landing page tests within paid social channels, offering a blend of theoretical knowledge and practical insights.
The Pearmill team will guide you through the nuances of setting up and managing landing page experiments on paid social platforms. You will learn about the critical rules to follow, the structure of effective tests, optimal conversion duration and budget allocation.
The session will also cover data analysis techniques and criteria for graduating landing pages.
In the second part of the webinar, Pearmill will explore the use of A/B testing platforms. Discover common pitfalls to avoid in A/B testing and gain insights into analyzing A/B tests results effectively.
Core Web Vitals SEO Workshop - improve your performance [pdf]Peter Mead
Core Web Vitals to improve your website performance for better SEO results with CWV.
CWV Topics include:
- Understanding the latest Core Web Vitals including the significance of LCP, INP and CLS + their impact on SEO
- Optimisation techniques from our experts on how to improve your CWV on platforms like WordPress and WP Engine
- The impact of user experience and SEO
5 big bets to drive growth in 2024 without one additional marketing dollar AND how to adapt to the biggest shifting eCommerce trend- AI.
1) Romance Your Customers - Retention
2) ‘Alternative’ Lead Gen - Advocacy
3) The Beautiful Basics - Conversion Rate Optimization
4) Land that Bottom Line - Profitability
5) Roll the Dice - New Business Models
Digital marketing is the art and science of promoting products or services using digital channels to reach and engage with potential customers. It encompasses a wide range of online tactics and strategies aimed at increasing brand visibility, driving website traffic, generating leads, and ultimately, converting those leads into customers.
https://nidmindia.com/
10 Video Ideas Any Business Can Make RIGHT NOW!
You'll never draw a blank again on what kind of video to make for your business. Go beyond the basic categories and truly reimagine a brand new advanced way to brainstorm video content creation. During this masterclass you'll be challenged to think creatively and outside of the box and view your videos through lenses you may have never thought of previously. It's guaranteed that you'll leave with more than 10 video ideas, but I like to under-promise and over-deliver. Don't miss this session.
Key Takeaways:
How to use the Video Matrix
How to use additional "Lenses"
Where to source original video ideas
The digital marketing industry is changing faster than ever and those who don’t adapt with the times are losing market share. Where should marketers be focusing their efforts? What strategies are the experts seeing get the best results? Get up-to-speed with the latest industry insights, trends and predictions for the future in this panel discussion with some leading digital marketing experts.
Most small businesses struggle to see marketing results. In this session, we will eliminate any confusion about what to do next, solving your marketing problems so your business can thrive. You’ll learn how to create a foundational marketing OS (operating system) based on neuroscience and backed by real-world results. You’ll be taught how to develop deep customer connections, and how to have your CRM dynamically segment and sell at any stage in the customer’s journey. By the end of the session, you’ll remove confusion and chaos and replace it with clarity and confidence for long-term marketing success.
Key Takeaways:
• Uncover the power of a foundational marketing system that dynamically communicates with prospects and customers on autopilot.
• Harness neuroscience and Tribal Alignment to transform your communication strategies, turning potential clients into fans and those fans into loyal customers.
• Discover the art of automated segmentation, pinpointing your most lucrative customers and identifying the optimal moments for successful conversions.
• Streamline your business with a content production plan that eliminates guesswork, wasted time, and money.
Mastering Multi-Touchpoint Content Strategy: Navigate Fragmented User JourneysSearch Engine Journal
Digital platforms are constantly multiplying, and with that, user engagement is becoming more intricate and fragmented.
So how do you effectively navigate distributing and tailoring your content across these various touchpoints?
Watch this webinar as we dive into the evolving landscape of content strategy tailored for today's fragmented user journeys. Understanding how to deliver your content to your users is more crucial than ever, and we’ll provide actionable tips for navigating these intricate challenges.
You’ll learn:
- How today’s users engage with content across various channels and devices.
- The latest methodologies for identifying and addressing content gaps to keep your content strategy proactive and relevant.
- What digital shelf space is and how your content strategy needs to pivot.
With Wayne Cichanski, we’ll explore innovative strategies to map out and meet the diverse needs of your audience, ensuring every piece of content resonates and connects, regardless of where or how it is consumed.
31. One model per answer option
One question per brand, different
competitor brands per market
~500 - 1500 models per project
32. One model per answer option
One question per brand, different
competitor brands per market
~500 - 1500 models per project
~20min per model
= ~10 days compute time
33. Integrates with AWS Batch
Easy to use for Data Scientists
Supports reproducibility
34. MRP as Metaflow
from metaflow import FlowSpec,step
class MRPFlow(FlowSpec):
@step
def start(self):
Self.data, self.questions = load_data()
self.next(self.run_model, foreach="questions")
@step
def run_model(self):
question = self.input
self.result = run_mrp(question, self.data)
self.next
@step
def join(self, inputs):
for result in inputs:
save(result)
self.next(self.end)
@step
def end(self):
pass
35. MRP as Metaflow
from metaflow import FlowSpec,step
class MRPFlow(FlowSpec):
@step
def start(self):
Self.data, self.questions = load_data()
self.next(self.run_model,
foreach="questions")
@step
def run_model(self):
question = self.input
self.result = run_mrp(question, self.data)
self.next
@step
def join(self, inputs):
for result in inputs:
save(result)
self.next(self.end)
@step
def end(self):
36. MRP as Metaflow
from metaflow import FlowSpec,step
class MRPFlow(FlowSpec):
@step
def start(self):
Self.data, self.questions = load_data()
self.next(self.run_model,
foreach="questions")
@step
def run_model(self):
question = self.input
self.result = run_mrp(question, self.data)
self.next
@step
def join(self, inputs):
for result in inputs:
save(result)
self.next(self.end)
@step
def end(self):
40. Challenges
Convergence
How to monitor convergence of
1000+ models?
More predictor variables
Full joint distribution needed of all
predictor variables.
41. Summary
● Multilevel-regression improves errors by using grouped structure
● Propagation of uncertainty improves weighting
Corrie Bartelheimer
Senior Data Scientist
corrie.bartelheimer@latana.com
42. Introductory book on Bayesian Statistics: https://xcelab.net/rm/statistical-rethinking/
Stan: https://mc-stan.org/
Stan interface brms (R): https://paul-buerkner.github.io/brms/
MRP: Forecasting elections with non-representative polls https://www.sciencedirect.com/science/article/abs/pii/S0169207014000879
Metaflow https://metaflow.org/
MRP at Latana:
- https://latana.com/whitepapers/mrp-vs-traditional-quota-sampling-brand-tracking/
- https://aws.amazon.com/blogs/startups/brand-tracking-with-bayesian-statistics-and-aws-batch/
Resources and
Links
Editor's Notes
What is brand tracking? A company/brand is interested in how many people are aware of their brand (“brand awareness”).
Brands are usually interested in the following questions: How is their brand faring with specific target groups (e.g. they might only be interested in women, or only in people living in big cities).
Another question is how people perceive their brand, what do they associate with it (is it fun, or intuitive, etc…).
Of course, one important question is also if there have been changes over time. This is also relevant if the brand wants to know if their marketing campaign was successful and had an impact.
It’s easy to get a bunch of people to respond to an online survey but any target group (eg. women with kids) will most likely be small.
This can make it hard (to impossible) to say if any change is due to a real change in how consumers see the brand or if it’s just a random change in the respondents composition.
Also, online surveys are usually not very representative of the general population.
Traditional approaches (and their shortcomings)
One approach is to collect respondents data and weight the different groups afterwards. Let’s say we’re interested in both gender and the age.
We can then use census data to determine weights for each demographic cell. For each cell, we estimate the mean from the respondents data, multiply with the weight and sum up to get an estimate of the general population.
Imagine though we get data like this. Now the estimate based on only two respondents has the biggest weight and thus gets amplified. This means that small changes in this group can lead to big changes in the overall outcome. If we’re not interested in the overall estimate for the general population but for one of the subgroups, then (especially for the small groups) we get such large errors that the result is basically unusable.
Quota sampling goes the other way around and sets numbers of how many responses should be collected from each group.
The problem with quota sampling however, is that it can take a while to fill all cells with the number of responses needed. This also makes it more expensive.
(Also possible to stop early with collecting data and then weight results).
Paper: see e.g. Forecasting elections with non-representative polls https://www.sciencedirect.com/science/article/abs/pii/S0169207014000879
First part: Multilevel Regression
The “Bayesian part” of the method, a hierarchical model. Our variable of interest “knows brand” is modelled by a Bernoulli likelihood
In our example from the previous slides, our predictor variables are gender and age.
We model gender and age as random intercepts/hierarchical.
Since it’s a Bayesian model, we also use priors of course.
Some intuition: for the gender parameter, this means that we estimate two parameters: male and female.
But since we group the two parameters, both come from a common distribution. This means that we allow both parameter to be different but we also think they should be similar.
If there is little data for one group, then the model will estimate the parameter as close to the estimate for all related groups.
In our example this means that any demographic cell borrows strength from neighboring cells (similar groups).
The estimate for ♂, 30-60 is then the sum of the parameter for ♂ and the parameter 30-60. Both parameters also take into account the information from the other groups. Thus, each cell is not considered in isolation but we also use information from neighboring cells (similar groups)
Second part: Poststratification (weighting)
The poststratification part is basically the weighting method.
To get an estimate for the general popluation, we compute the prediction from our model for the different cells (= proportion of men/women that know brand) and weight this with the proportion of men/women in the general population.
Mathematically, the weight corresponds to the probability of gender and the prediction to the probability of knows brand conditioned on gender. Combining them gives the joint distribution pr(knows_brand, gender) according to the chain rule. Adding them up is equivalent to marginalizing out gender.
The difference to the approach outlines before, is that the predictions are actually samples from our posterior instead of point estimates. Using a Bayesian model thus allows us to propagate the uncertainty.
Back to our use case, on top of questions asking for “do you know this brand?” we also have questions like these.
These questions are binarized, thus resulting in one model per option (here 6)
The same question is also asked for competitor brands, there are different competitors for different markets which means we easily end up with around 1000 models even though just a handful of questions were asked.
Each model takes around 20min, so in total we have 10 days of computing time.
So many models. So little time.
Metaflow Library: A library developed by Netflix https://metaflow.org/
It integrates nicely with AWS Batch, is easy to use even for Data Scientists that don’t have strong background in cloud computing and additionally also supports reproducibility.
This is how MRP could look like in Metaflow.
It’s a DAG: Each step represents a job launched on Batch, where the model step is parallelized, so one job per model is launched.
The heavy lifting here is done using the “foreach” keyword. If questions is a list then with the foreach, metaflow will start one run_model job for each question. All the orchestration is being taken care of by metaflow.
To increase resources, we can add just line
Some remaining challenges: how to monitor convergence of more than 1000 models?
How to add more predictor variables, especially custom variables specific to one brand for which no census data is available
