As AI-powered tools become increasingly prevalent, prompt engineering is becoming a skill that developers need to master. Large language models (LLMs) and other generative foundation models require contextual, specific, and tailored natural language instructions to generate the desired output. This means that developers need to write prompts that are clear, concise, and informative. 

In this blog, we will explore six best practices that will make you a more efficient prompt engineer. By following our advice, you can begin creating more personalized, accurate, and contextually aware applications. So let’s get started!

Tip #1: Know the model’s strengths and weaknesses

As AI models evolve and become more complex, it is essential for developers to comprehend their capabilities and limitations. Understanding these strengths and weaknesses can help you, as a developer, avoid making mistakes and create safer, more reliable applications.

For example, an AI model that is trained to recognize images of blueberries may not be able to recognize images of strawberries. Why? Because the model was only trained on a dataset of blueberry images. If a developer uses this model to build an application that is supposed to recognize both blueberries and strawberries, the application would likely make mistakes, leading to an ineffective outcome, and poor user experience.

It’s important to note that AI models have the ability to be biased. This is due to AI models being trained on data that is collected from the real world, and so it can reflect the inequitable power dynamics inherently rooted in our social hierarchy. If the data that is used to train an AI model is biased, then the model will also be biased. This can lead to problems if the model is used to make decisions that affect people by reinforcing societal biases. Addressing these biases is important to ensure that data is fair, promoting equality, and ensuring the responsibility of AI technology. Prompt engineers should be aware of training limitations or biases so they can craft prompts more effectively and understand what kind of prompting is even possible for a given model.

Tip #2: Be as specific as possible

AI models have the ability to comprehend a variety of prompts. For instance Google’s PaLM 2 can understand natural language prompts, multilingual text, and even programming codes like Python and JavaScript. Although AI models can be very knowledgeable, they are still imperfect, and have the ability to misinterpret prompts that are not specific enough. In order for AI models to navigate ambiguity, it is important to tailor your prompts specifically to your desired outcome. 

Let’s say you would like your AI model to generate a recipe for 50 vegan blueberry muffins. If you prompt the model with “what is a recipe for blueberry muffins?”, the model does not know that you need to make 50 muffins. It is thus unlikely to list the larger volume of ingredients you’ll need or include tips to help you more efficiently bake such a large number of muffins. The model can only go off the context that is provided. A more effective prompt would be “I am hosting 50 guests. Generate a recipe for 50 blueberry muffins.” The model is more likely to generate a response that is relevant to your request and meets your specific requirements.

Tip #3: Utilize contextual prompts

Utilize contextual information in your prompts to help the model gain an in-depth understanding of your requests. Contextual prompts can include the specific task you want the model to perform, a replica of the output you’re looking for, or a persona to emulate, from a marketer or engineer to a high school teacher. Defining a tone and perspective for an AI model gives it a blueprint of the tone, style, and focused expertise you’re looking for to improve the quality, relevance, and effectiveness of your output. 

In the case of the blueberry muffins, it is important to prompt the model using the context of the situation. The model might need more context than generating a recipe for 50 people. If it needs to be aware that the recipe must be vegan friendly, you might prompt the model by asking it to answer by emulating a skilled vegan chef. 

By providing contextual prompts, you can help ensure that your AI interactions are as seamless and efficient as possible. The model will be able to more quickly understand your request and it will be able to generate more accurate and relevant responses.

Tip #4: Provide AI models with examples 

When creating prompts for AI models, it is helpful to provide examples. This is because prompts act as instructions for the model, and examples can help the model to understand what you are asking for. Providing a prompt with an example looks something like this: “here are several recipes I like – create a new recipe based on the ones I provided.” The model can now understand the your ability and needs in order to make this pastry,

Tip #5: Experiment with prompts and personas

The way you construct your prompt impacts the model’s output. By creatively exploring different requests, you will soon have an understanding of how the model weighs its answers, and what happens when you interfuse your domain knowledge, expertise, and lived experience with the power of a multi-billion parameter large language model. 

Try experimenting with different keywords, sentence structures, and prompt lengths to discover the perfect formula. Allow yourself to step into the shoes of various personas, from work personas such as “product engineer” or “customer service representatives,” to parental figures or celebrities such as your grandmother, a celebrity chef, and explore everything from cooking to coding!

By crafting unique, and innovative, requests replete with your expertise and experience, you can learn which prompts provide you with your ideal output. Further refining your prompts, known as ‘tuning,’ allows the model to have a greater understanding and framework for your next output.

Tip #6: Try chain-of-thought prompting

Chain of thought prompting is a technique for improving the reasoning capabilities of large language models (LLMs). It works by breaking down a complex problem into smaller steps, and then prompting the LLM to provide intermediate reasoning for each step. This helps the LLM to understand the problem more deeply, and to generate more accurate and informative answers. This will help you to understand the answer better and to make sure that the LLM is actually understanding the problem. 

Conclusion

Prompt engineering is a skill that all workers, across industries and organizations, will need as AI-powered tools are becoming more prevalent. Remember to incorporate these five essential tips the next time you communicate with an AI model, so you can generate the accurate outputs that you desire. AI will forever continue to develop, constantly refining itself as we use it, so I encourage you to remember that learning, for mind and machine, is a never ending journey. Happy Prompting!

We are thrilled to announce that Google has been named a Leader in The Forrester Wave™: Cloud Data Warehouse, Q1 2021 report. For more than a decade, BigQuery, our petabyte-scale cloud data warehouse, has been in a class of its own. We’re excited to share this recognition and we want to thank our strong community of customers and partners for voicing their opinion. We believe this report validates the alignment of our strategy with our customers’ analytics needs.

“Customers like Google’s frequency of data warehouse releases, business value, future proof architecture, high-end scale, geospatial capabilities, strong AI/ML capabilities, good security capabilities, and broad analytical use cases,” according to the Forrester report. Today’s data leaders require a data warehousing platform that provides both depth and breadth and with BigQuery, organizations are able to unlock deeper data science and machine learning capabilities while promoting data democratization and providing the highest levels of availability. 

Google BigQuery: 5 out of 5!

Forrester gave Google BigQuery a score of 5 out of 5 across 19 different criteria, including:

Today, customers across the globe use BigQuery to run business critical analytics workloads to enable BI acceleration, IoT analytics, customer intelligence, AI/ML-based analytics, data science, data collaboration, and data services. Customers such as Verizon, Wayfair, HSBC, Twitter, AirAsia, KeyBank, The Home Depot, and Vodafone have anchored their digital transformation efforts on BigQuery—unlocking deeper insights for their people. 

Customers use BigQuery across all industries, to solve issues like credit card fraud detection, predictive forecasting, anomaly detection, log analytics and many more. Forrester recognized this work and gave BigQuery a 5/5 score for supporting vertical and horizontal use cases. 

More BigQuery advantages 

Google is the first hyperscale provider to offer a multi cloud data analytics solution. With BigQuery Omni, customers can perform cross-cloud analytics with ease, and drive business outcomes they couldn’t achieve with the siloed approach offered by other vendors. 

We designed BigQuery to be highly scalable and more open and interoperable so that customers can join data across SQL databases, traditional unstructured data lakes in object storage, and even spreadsheets using any analysis tool. 

Recent innovations like BigQuery BI Engine lets us provide the best analytics experience by delivering sub-second query response times from any business intelligence tool, from Google’s Looker and Connected Sheets to Tableau, Microsoft Power BI, ThoughtSpot, and others. Our goal is to meet customers where they are rather than force them into a one-size-fits-all approach to data analysis. 

With BigQuery, organizations gain both breadth and depth of capabilities to transform their analytics strategy. Forrester also gave BigQuery 5 out of 5 in:

Data-powered innovation 

These advantages enable our customers to accelerate their digital transformation and reimagine their business through data-powered innovation. Google’s leadership across AI, analytics, and databases comes together in a single data cloud platform that provides everyone with the ability to get value out of their data faster.

We bring decades of research and innovation in AI to our customers through industry-leading AI solutions, that in turn helps our customers solve their biggest problems. Our support of open standards and APIs enables interoperability between a variety of services for ingestion, storage, processing and analytics across the data cloud platform. And finally, we believe our multi-layered security approach throughout the data stack ensures redundancy and reliability so that customers can have the peace of mind that their data is always protected. 

We are honored to be a leader in this Forrester Wave™ and look forward to continuing to innovate and partner with you on your digital transformation journey. 

Download the full Forrester Wave™ :Cloud Data Warehouse, Q1 2021 report. And check out these smart analytics reference patterns. To learn more about BigQuery, visit our website, and get started immediately with the free BigQuery Sandbox.

Editor’s note: This blog has been updated since it originally published to reflect the total number of generative AI courses that have become available since the publish date.

An AI/ML (artificial intelligence/machine learning) career path can be a great specialty area within the cloud—and one of the most accessible! Because this area is constantly developing, most recently with the rise of generative AI, I want to share some recommendations to help you chart a sustainable career as AI/ML continues to evolve.

Generative AI falls under the overall category of AI and offers a new and exciting way of interacting with information, brands, and other people. Let’s talk about how these disciplines work together and about materials available to help you upskill in these areas.

To this end, we are happy to announce a new set of generative AI training content available at no cost. So, whether you are just getting started or already have a more advanced role, read on to find ways to help reach your desired position. 

ML compared to AI

While the terms ML and AI are often used interchangeably, there are distinct differences. AI is an umbrella concept wherein machines are taught to perform tasks normally associated with human intelligence, such as decision-making and language interaction. ML is a subset of AI dedicated to taking data from the past and training algorithms to create models that can perform highly complex tasks without being explicitly programmed. It’s the basis for most forms of AI that people interact with, like virtual assistants, music recommendations, and chatbots.

Data engineers and ML engineers work with data scientists to get insights from data. They are needed to create software models and get clear results, and develop deployable applications. These skilled roles are needed in every industry!

Change the world for the better with generative AI

I recently wrote about four key pillars of technology trends expected in the next decade, including the role of AI/ML in the cloud environment as one of those pillars, and how you can bridge the skills gap and build your career.

Generative AI is a new type of ML that has made a lot of headlines recently. Research from CIO Dive finds that seven in 10 executives say their companies are investigating or exploring generative AI. Now is a great time to become an expert, while we are at the cusp of this technology becoming more widely adopted. 

To get to generative AI, we need to talk first about deep learning. Deep learning is a subset of ML that uses artificial neural networks to process more complex patterns than traditional ML. Generative AI sits still further down the funnel, as a subset of deep learning that typically involves the Transformer architecture. Essentially, it’s a type of AI that can map long-range dependencies and patterns in large training sets, then use what it learns to produce new content, including text, imagery, audio, and synthetic data.

Generative AI relies on large models, such as large language models (LLMs) that can classify and generate text, answer questions, and summarize documents. For more detail about how this works, check out the video below from Google I/O 2023, which features great information from Dr. Gwendolyn Stripling, Artificial Intelligence Technical Curriculum Developer for Google Cloud.

How Google is offering generative AI

Here at Google, we’ve been heavily invested for decades in research and innovation aimed at helping businesses, governments, and developers maximize the potential of AI. Our vision is to empower builders, innovators, developers, and doers to use AI in unique, responsible, productive ways. 

To deliver on this, we recently announced a variety of solutions to bring generative AI into our offerings, beginning with Generative AI support on Vertex AI and Gen App Builder. 

Vertex AI is Google Cloud’s ML platform for training and deploying ML models and AI applications. With Generative AI support on Vertex AI, data science teams and developers can access foundation models from both Google and other sources, helping them to quickly build, customize, and deploy models for their own use cases. 

As part of generative AI support on Vertex AI, Model Garden and Generative AI Studio are now in preview, including access to PaLM 2 for Text and Chat and Embeddings API for Text, while other features and services are available to select trusted testers. If you’d like early access to Google Cloud’s AI products, join the waitlist here.

We announced our latest PaLM model in production at I/O 2023: PaLM 2. Today, it powers more than 25 products! Additionally, we’ve fine-tuned PaLM 2 for specific use cases, including security and medical domains. You’ve probably also heard of Bard, our experiment for conversational AI, which is fully running on PaLM 2. 

Our Gen App Builder aims to make customer, partner, and employee interactions more effective and helpful. With it, developers — even those with limited data science expertise — can quickly create bots, chat apps, digital assistants, custom search engines, and more. Gen App Builder even makes it possible to create some generative AI apps without any coding skills. 

Now, let’s bring these products to life! Not sure exactly the use case or in need of a little inspiration? Check out this blog from Google Cloud CEO Thomas Kurian to see how customers and partners in the ecosystem are bringing ideas to life with generative AI. And watch this video to see generative AI in action!

Build your AI/ML skills and validate your knowledge to grow your career

There is a lot of excitement around generative AI—it’s truly a brand new path to consider for your AI/ML career! Check out my recommendations below for training options for AI/ML roles. These will help you gain critical skills as generative AI becomes more widely available.

*NEW* training materials, specific to generative AI technologies
Generative AI Learning Path – no cost training 

• [Course] Introduction to Generative AI (1 day)
• [Course] Introduction to Large Language Models (1 day)
• [Course] Attention Mechanism (1 day)
• [Course] Transformer Models and BERT Model (1 day)
• [Course] Introduction to Image Generation (1 day)
• [Course] Create Image Captioning Models (1 day)
• [Course] Encoder-Decoder Architecture (1 day)

Additional AI/ML training – varying learning credits required to complete on Google Cloud Skills Boost 
Introductory level

• [Course with completion badge] How Google Does Machine Learning (1 day)
• [Course with completion badge] MLOps: Getting Started (1 day)
• [Skill badge] Get started with TensorFlow on Google Cloud (8 hours)  
• [Skill badge] Perform foundational ML, AI and data tasks in Google Cloud (7 hours)
• [Course with completion badge] Language, Speech, Text, and Translation with Google Cloud APIs (5 hours)

Intermediate/Multi-level

• [Learning Path] Machine Learning Engineer (Collection of 15 video courses and labs) 
• [Skill badge] Build and Deploy Machine Learning Solutions on Vertex AI (1 day)

Advanced

• [Training + certification exam] Google Cloud Professional ML Engineer Certification (varied time)
• [Skill badge] Machine Learning with TensorFlow in Vertex AI (90 minutes)
• [Course with completion badge] Natural Language Processing in Google Cloud (1 day)

You can also catch up on hands-on demos in our most recent Cloud OnBoard: From Data to AI with BigQuery and Vertex AI. In it, we guide you through all steps of the data-to-AI solution.

Coming up on June 29, we are hosting Getting Started with Vertex Generative AI as part of Innovators Live. Register now!

I also encourage you to consider an Innovators Plus subscription. You get access to over 700 labs, skill badges and courses (including many that are specific to AI/ML), a certification exam voucher, and up to $1,000 in Google Cloud credits. You’ll also get access to live learning events and technical briefings with Google Cloud experts, plus 1:1 consultations (talk with us about your upcoming generative AI project!). It’s a $1,500 package for only $299. Learn more!

Clinical Decision Support System (CDSS) is an important technology for the healthcare industry that analyzes data to help healthcare professionals make decisions related to patient care. The market size for the global clinical decision support system appears poised for expansion, with one study predicting a compound annual growth rate (CAGR) of 10.4%, from 2022 to 2030, to $10.7 billion.

For any health organization that wants to build a CDSS system, one key block is to locate and extract the medical entities that are present in the clinical notes, medical journals, discharge summaries, etc. Along with entity extraction, the other key components of the CDSS system are capturing the temporal relationships, subjects, and certainty assessments.

At Google Cloud, we know how critical it is for the healthcare industry to build CDSS systems, so we worked with Apollo 24|7, the largest multi-channel digital healthcare platform in India, to build the key blocks of their CDSS solution.

We helped them to parse the discharge summaries and prescriptions to extract the medical entities. These entities can then be used to build a recommendation engine that would help doctors with the “Next Best Action” recommendation for medicines, lab tests, etc.

Let’s take a sneak peek at Apollo 24|7’s entity extraction solutions, and the various Google AI technologies that were tested to form the technology stack.

Datasets Used

To perform our experiments on entity extraction, we used two types of datasets.

1. i2b2 Dataset – i2b2 is an open-source clinical data warehousing and analytics research platform that provides annotated deidentified patient discharge summaries made available to the community for research purposes. This dataset was primarily used for training and validation of the models.
2. Apollo 24|7’s Dataset – De-identified doctor’s notes from Apollo24|7 were used for testing. Doctors annotated them to label the entities and offset values.

Experimentation and choosing the right approach — Four models put to test

For entity extraction, both Google Cloud products and open-source approaches were explored. Below are the details:

1. Healthcare Natural Language API: This is a no-code approach that provides machine learning solutions for deriving insights from medical text. Using this, we parsed unstructured medical text and then generated a structured data representation of the medical knowledge entities stored in the data for downstream analysis and automation. The process includes:

• Extract information about medical concepts like diseases, medications, medical devices, procedures, and their clinically relevant attributes;
• Map medical concepts to standard medical vocabularies such as RxNorm, ICD-10, MeSH, and SNOMED CT (US users only);
• Derive medical insights from text and integrate them with data analytics products in Google Cloud.

The advantage of using this approach is that it not only extracts a wide range of entity types like MED_DOSE, MED_DURATION, LAB_UNIT, LAB_VALUE, etc, but also captures functional features such as temporal relationships, subjects, and certainty assessments, along with the confidence scores. Since it is available on Google Cloud, this offers long-term product support. It is also the only fully-managed NLP service among all the approaches tested and hence, it requires the least effort to implement and manage.

But one thing to keep in mind is that since the Healthcare NL API offers natural language models that are pre-trained, it currently cannot be used for custom entity extraction models trained using custom annotated medical text or to extract custom entities. This has to be done via AutoML Entity Extraction for Healthcare, another Google Cloud service for custom model development. Custom model development is important for adapting the pre-trained models to new languages or region-specific natural language processing, such as medical terms whose use may be more prevalent in India than in other regions

2. Vertex AutoML Entity Extraction for Healthcare: This is a low-code approach that’s already available on Google Cloud. We used AutoML Entity Extraction to build and deploy custom machine learning models that analyzed documents, categorized them, and identified entities within them. This custom machine learning model was trained on the annotated dataset provided by the Apollo 24|7 team.

The advantage of AutoML Entity Extraction is that it gives the option to train on a new dataset. However, one of the prerequisites to keep in mind is that it needs a little pre-processing to capture the input data in the required JSONL format. Since this is an AutoML model just for Entity Extraction, it does not extract relationships, certainty assessments, etc.

3. BERT-based Models on Vertex AI: Vertex AI is Google Cloud’s fully managed unified AI platform to build, deploy, and scale ML models faster, with pre-trained and custom tooling. We experimented with multiple custom approaches based on pre-trained BERT-based models, which have shown state-of-the-art performance in many natural language tasks. To gain better contextual understanding of medical terms and procedures, these BERT-based approaches are explicitly trained on medical domain data. Our experiments were based on BioClinical BERT, BioLink BERT, Blue BERT trained on Pubmed dataset, and Blue BERT trained on Pubmed + MIMIC datasets.

The major advantage of these BERT-based models is that they can be finetuned on any Entity Recognition task with minimal efforts.

However, since this is a custom approach, it requires some technical expertise. Additionally, it does not extract relationships, certainty assessments, etc. This is one of the main limitations of using BERT-based models.

4. ScispaCy on Vertex AI: We used Vertex AI to perform experiments based on ScispaCy, which is a Python package containing spaCy models for processing biomedical, scientific or clinical text.

Along with Entity Extraction, Scispacy on Vertex AI provides additional components like Abbreviation Detector, Entity Linking, etc. However, when compared to other models, it was less precise, with too many junk phrases, like “Admission Date,” captured as entities.

“Exploring multiple approaches and understanding the pros/cons of each approach helped us to decide the one that would fit our business requirements.” according to Abdussamad M, Engineering Lead at Apollo 24|7.

Evaluation Strategy

In order to match the parsed entity with the test data labels, we used extensive matching logic that comprised of the below four methods:

1. Exact Match – Exact match captures entities where the model output and the entities in the test dataset match. Here, the offset values of the entities have also been considered. For example, the entity “gastrointestinal infection” that is present as-is in both the model output and the test label will be considered an “Exact Match.”
2. Match-Score Logic – We used a scoring logic for matching the entities. For each word in the test data labels, every word in the model output is matched along with the offset. A score is calculated between the entities and based on the threshold, it is considered as a match.
3. Partial Match – In this matching logic, entities like “hypertension” and “hypertensive” are matched based on the Fuzzy logic.
4. UMLS Abbreviation Lookup – We also observed that the medical text had some abbreviations, like AP meaning abdominal pain. These were first expanded by doing a lookup on the respective UMLS (Unified Medical Language System) tables and then passed to the individual entity extraction models.

Performance Metrics

We used precision and recall metrics to compare the outcomes of different models/experiments.

Precision (also called positive predictive value) is the fraction of relevant instances among the retrieved instances, while recall (also known as sensitivity) is the fraction of relevant instances that were retrieved.

The below example shows how to calculate these metrics for a given sample.

Example sample: “Krish has fever, headache and feels uncomfortable”

Expected Entities: [“fever”, “headache”]

Model Output: [“fever”, “feels”, “uncomfortable”]

Thus,

Experimentation Results
The following table captures the results of the above experiments on Apollo24|7’s internal datasets.

Finally, the Blue BERT model trained on the Pubmed dataset had the best performance metrics with a 81% improvement on Apollo 24|7’s baseline mode with the Healthcare Natural Language API providing the context, relationships, and codes. This performance could be further improved by implementing an ensemble of these two models.

“With the Blue BERT model giving the best performance for entity extraction on Vertex AI and the Healthcare NL API being able to extract the relationships, certainty assessments etc, we finally decided to go with an ensemble of these 2 approaches,“ Abdussamad added.

Fast track end-to-end deployment with Google Cloud AI Services (AIS)

Google AIS (Professional Services Organization) helped Apollo24|7 to build the key blocks of the CDSS system.

The partnership between Google Cloud and Apollo 24|7 is just one of the latest examples of how we’re providing AI-powered solutions to solve complex problems to help organizations drive the desired outcomes. To learn more about Google Cloud’s AI services, visit our AI & ML Products page, and to learn more about Google Cloud solutions for health care, explore our Google Cloud Healthcare Data Engine page.

Acknowledgements

We’d like to give special thanks to Nitin Aggarwal, Gopala Dhar and Kartik Chaudhary for their support and guidance throughout the project. We are also thankful to Manisha Yadav, Santosh Gadgei and Vasantha Kumar for implementing the GCP infrastructure. We are grateful to the Apollo team (Chaitanya Bharadwaj, Abdussamad GM, Lavish M, Dinesh Singamsetty, Anmol Singh and Prithwiraj) and our partner team from HCL/Wipro (Durga Tulluru and Praful Turanur) who partnered with us in delivering this successful project. Special thanks to the Cloud Healthcare NLP API team (Donny Cheung, Amirhossein Simjour, and Kalyan Pamarthy).