Add Acrolinx suggestions

Author: psanapathi

articles/azure-health-insights/oncophenotype/transparency-note.md

@@ -9,9 +9,9 @@ ms.date: 04/11/2023
 ms.author: prsanapa
 ---
 
-# Transparency note for Onco Phenotype
+# Transparency Note for Onco Phenotype
 
-## What is a transparency note?
+## What is a Transparency Note?
 
 An AI system includes not only the technology, but also the people who will use it, the people who will be affected by it, and the environment in which it is deployed. Creating a system that is fit for its intended purpose requires an understanding of how the technology works, what its capabilities and limitations are, and how to achieve the best performance. Microsoft’s Transparency Notes are intended to help you understand how our AI technology works, the choices system owners can make that influence system performance and behavior, and the importance of thinking about the whole system, including the technology, the people, and the environment. You can use Transparency Notes when developing or deploying your own system, or share them with the people who will use or be affected by your system.
 
@@ -53,11 +53,11 @@ The Onco Phenotype model can be used in the following  scenario. The system’s
 
 #### Considerations when choosing a use case
 
-We encourage customers to leverage the Onco Phenotype model in their innovative solutions or applications. However, here are some considerations when choosing a use case:
+We encourage customers to use the Onco Phenotype model in their innovative solutions or applications. However, here are some considerations when choosing a use case:
 
-- **Avoid scenarios that use personal health information for a purpose not permitted by patient consent or applicable law.** Health information has special protections regarding privacy and consent. Make sure that all data you use has patient consent for the way you use the data in your system or you are otherwise compliant with applicable law as it relates to the use of health information.
-- **Facilitate human review and inference error corrections.** Given the sensitive nature of health information, it is essential that a human review the source data and correct any inference errors.
-- **Avoid scenarios that use this service as a medical device, for clinical support, or as a diagnostic tool or workflow without a human in the loop.** The system was not designed for use as a medical device, for clinical support, or as a diagnostic tool for the diagnosis, cure, mitigation, treatment, or prevention of disease or other conditions without human intervention. A qualified professional should always verify the inferences and relevant evidence before finalizing or relying on the information.
+- **Avoid scenarios that use personal health information for a purpose not permitted by patient consent or applicable law.** Health information has special protections regarding privacy and consent. Make sure that all data you use has patient consent for the way you use the data in your system or you're otherwise compliant with applicable law as it relates to the use of health information.
+- **Facilitate human review and inference error corrections.** Given the sensitive nature of health information, it's essential that a human review the source data and correct any inference errors.
+- **Avoid scenarios that use this service as a medical device, for clinical support, or as a diagnostic tool or workflow without a human in the loop.** The system wasn't designed for use as a medical device, for clinical support, or as a diagnostic tool for the diagnosis, cure, mitigation, treatment, or prevention of disease or other conditions without human intervention. A qualified professional should always verify the inferences and relevant evidence before finalizing or relying on the information.
 
 ## Limitations
 
@@ -66,8 +66,8 @@ We encourage customers to leverage the Onco Phenotype model in their innovative
 Specific characteristics and limitations of the Onco Phenotype model include:
 
 - **Multiple cancer cases for a patient:** The model infers only a single set of phenotype values (tumor site, histology, and clinical/pathologic stage TNM categories) per patient. If the model is given an input with multiple primary cancer diagnoses, the behavior is undefined and might mix elements from the separate diagnoses.
-- **Inference values for tumor site and histology:** The inference values are only as exhaustive as the training dataset labels. If the model is presented with a cancer case for which the true tumor site or histology was not encountered during training (for example, a rare tumor site or histology), the model will be unable to produce a correct inference result.
-- **Clinical/pathologic stage (TNM categories):** The model does not currently identify the initiation of a patient's definitive treatment. Therefore, it might use clinical stage evidence to infer a pathologic stage value or vice-versa. Manual review should verify that appropriate evidence supports clinical and pathologic stage results. The model does not predict subcategories or isolated tumor cell modifiers. For instance, T3a would be predicted as T3, and N0(i+) would be predicted as N0.
+- **Inference values for tumor site and histology:** The inference values are only as exhaustive as the training dataset labels. If the model is presented with a cancer case for which the true tumor site or histology wasn't encountered during training (for example, a rare tumor site or histology), the model will be unable to produce a correct inference result.
+- **Clinical/pathologic stage (TNM categories):** The model doesn't currently identify the initiation of a patient's definitive treatment. Therefore, it might use clinical stage evidence to infer a pathologic stage value or vice-versa. Manual review should verify that appropriate evidence supports clinical and pathologic stage results. The model doesn't predict subcategories or isolated tumor cell modifiers. For instance, T3a would be predicted as T3, and N0(i+) would be predicted as N0.
 
 ## System performance
 
@@ -76,13 +76,13 @@ In many AI systems, performance is often defined in relation to accuracy or by
 | **Outcome**    | **Correct/Incorrect** | **Definition** | **Example** |
 | -------------- | --------------------- | -------------- | ----------- |
 | True Positive  | Correct   | The system returns the tumor site as lung and that would be expected from a human judge. | The system correctly infers the tumor site as lung on the clinical documents of a lung cancer patient. |
-| True Negative  | Correct   | The system does not return the tumor site as lung, and this aligns with what would be expected from a human judge. | The system returns the tumor site as breast on the clinical documents of a breast cancer patient. |
-| False Positive | Incorrect | The system returns the tumor site as lung where a human judge would not. | The system returns the tumor site as lung on the clinical documents of a breast cancer patient. |
-| False Negative | Incorrect | The system does not return the tumor site as lung where a human judge would identify it as lung. | The system returns the tumor site as breast on the clinical documents of a lung cancer patient. |
+| True Negative  | Correct   | The system doesn't return the tumor site as lung, and this aligns with what would be expected from a human judge. | The system returns the tumor site as breast on the clinical documents of a breast cancer patient. |
+| False Positive | Incorrect | The system returns the tumor site as lung where a human judge wouldn't. | The system returns the tumor site as lung on the clinical documents of a breast cancer patient. |
+| False Negative | Incorrect | The system doesn't return the tumor site as lung where a human judge would identify it as lung. | The system returns the tumor site as breast on the clinical documents of a lung cancer patient. |
 
 ### Best practices for improving system performance
 
-For each inference, the Onco Phenotype model returns a confidence score that expresses how confident the model is with the response. Confidence scores range from 0 to 1. The higher the confidence score, the more certain the model is about the inference value it provided. However, the system is not designed for workflows or scenarios without a human in the loop. Also, inference values cannot be consumed without human review, irrespective of the confidence score. You can choose to completely discard an inference value if its confidence score is below a confidence score threshold that best suits the scenario.
+For each inference, the Onco Phenotype model returns a confidence score that expresses how confident the model is with the response. Confidence scores range from 0 to 1. The higher the confidence score, the more certain the model is about the inference value it provided. However, the system isn't designed for workflows or scenarios without a human in the loop. Also, inference values can't be consumed without human review, irrespective of the confidence score. You can choose to completely discard an inference value if its confidence score is below a confidence score threshold that best suits the scenario.
 
 ## Evaluation of Onco Phenotype
 
@@ -100,7 +100,7 @@ Microsoft has also tested the generalizability of the model by evaluating the tr
 
 At Microsoft, we strive to empower every person on the planet to achieve more. An essential part of this goal is working to create technologies and products that are fair and inclusive. Fairness is a multi-dimensional, sociotechnical topic and impacts many different aspects of our product development. You can learn more about Microsoft’s approach to fairness [here](https://www.microsoft.com/ai/responsible-ai?rtc=1&activetab=pivot1:primaryr6).
 
-One dimension we need to consider is how well the system performs for different groups of people. This might include looking at the accuracy of the model and measuring the performance of the complete system. Research has shown that without conscious effort focused on improving performance for all groups, it is often possible for the performance of an AI system to vary across groups based on factors such as race, ethnicity, language, gender, and age.
+One dimension we need to consider is how well the system performs for different groups of people. This might include looking at the accuracy of the model and measuring the performance of the complete system. Research has shown that without conscious effort focused on improving performance for all groups, it's often possible for the performance of an AI system to vary across groups based on factors such as race, ethnicity, language, gender, and age.
 
 The evaluation performance of the Onco Phenotype model was stratified by race to ensure minimal performance discrepancy between different patient racial groups. The lowest performance by racial group is well within 80% of the highest performance by racial group. When the evaluation performance was stratified by gender, there was no significant difference.
 
@@ -115,7 +115,7 @@ When getting ready to integrate and use AI-powered products or features, the fol
 - **Understand what it can do:** Fully vet and review the capabilities of Onco Phenotype to understand its capabilities and limitations.
 - **Test with real, diverse data:** Understand how Onco Phenotype will perform in your scenario by thoroughly testing it by using real-life conditions and data that reflects the diversity in your users, geography, and deployment contexts. Small datasets, synthetic data, and tests that don't reflect your end-to-end scenario are unlikely to sufficiently represent your production performance.
 - **Respect an individual's right to privacy:** Collect data and information from individuals only for lawful and justifiable purposes. Use data and information that you have consent to use only for this purpose.
-- **Legal review:** Obtain appropriate legal advice to review your solution, particularly if you will use it in sensitive or high-risk applications. Understand what restrictions you might need to work within and your responsibility to resolve any issues that might come up in the future.
+- **Legal review:** Obtain appropriate legal advice to review your solution, particularly if you'll use it in sensitive or high-risk applications. Understand what restrictions you might need to work within and your responsibility to resolve any issues that might come up in the future.
 - **System review:** If you're planning to integrate and responsibly use an AI-powered product or feature in an existing system of software or in customer and organizational processes, take the time to understand how each part of your system will be affected. Consider how your AI solution aligns with Microsoft's Responsible AI principles.
 - **Human in the loop:** Keep a human in the loop. This means ensuring constant human oversight of the AI-powered product or feature and maintaining the role of humans in decision-making. Ensure that you can have real-time human intervention in the solution to prevent harm. This enables you to manage where the AI model doesn't perform as expected.
 - **Security:** Ensure that your solution is secure and that it has adequate controls to preserve the integrity of your content and prevent unauthorized access.

articles/azure-health-insights/responsible-ai/transparency-note.md

@@ -12,7 +12,7 @@ ms.author: behoorne
 
 # Transparency Note for Project Health Insights
 
-## What is a transparency note?
+## What is a Transparency Note?
 
 An AI system includes not only the technology, but also the people who will use it, the people who will be affected by it, and the environment in which it is deployed. Creating a system that is fit for its intended purpose requires an understanding of how the technology works, what its capabilities and limitations are, and how to achieve the best performance. Microsoft’s Transparency Notes are intended to help you understand how our AI technology works, the choices system owners can make that influence system performance and behavior, and the importance of thinking about the whole system, including the technology, the people, and the environment. You can use Transparency Notes when developing or deploying your own system, or share them with the people who will use or be affected by your system.
 
@@ -30,7 +30,7 @@ Project Health Insights is an Azure Applied AI Service built with the Azure Cogn
 
 Read the overview to get an introduction to each feature and review the example use cases. See the How-to guides and the API reference to understand more details about what each feature does and what gets returned by the system.
 
-This article contains basic guidelines for how to use Azure Cognitive Service for language features responsibly. Read the general information first and then jump to the specific article if you're using one of the features below.
+This article contains basic guidelines for how to use Azure Cognitive Service for language features responsibly. Read the general information first and then jump to the specific article if you're using one of the following features.
 
 - [Transparency note for Trial Matcher](../trial-matcher/transparency-note.md) 
 - [Transparency note for Onco Phenotype](../oncophenotype/transparency-note.md)