Epoch AI description_key edits

etl/steps/data/garden/artificial_intelligence/latest/epoch_aggregates_affiliation.meta.yml

@@ -24,6 +24,8 @@ tables:
         short_unit: ''
         description_short: Describes the sector (Industry, Academia, or Collaboration) where the authors of an AI system have their primary affiliations.
         description_key:
+          - The authors of the Epoch dataset have established a set of criteria to identify key AI systems, which they refer to as notable. These systems must demonstrate the ability to learn, show tangible experimental results, and contribute advancements that push the boundaries of existing AI technology. In terms of notability, the AI must have garnered extensive academic attention, evidenced by a high citation count, hold historical significance in the field, mark a substantial advancement in technology, or be implemented in a significant real-world context. Recognizing the difficulty in evaluating the impact of newer AI systems since 2020 due to less available data, the authors also employ subjective judgment in their selection process for recent developments.
+
           - Systems are classified as "Industry" when their authors have ties to private sector entities, "Academia" when the authors come from universities or scholarly institutions, and "Industry - Academia Collaboration" if a minimum of 30% of the authors represent each sector.
         description_processing: >
           Processing involved calculating total number of AI systems developed within each category of reseacher affiliation for each year.
@@ -47,6 +49,8 @@ tables:
         short_unit: ''
         description_short: Describes the sector (Industry, Academia, or Collaboration) where the authors of an AI system have their primary affiliations.
         description_key:
+          - The authors of the Epoch dataset have established a set of criteria to identify key AI systems, which they refer to as notable. These systems must demonstrate the ability to learn, show tangible experimental results, and contribute advancements that push the boundaries of existing AI technology. In terms of notability, the AI must have garnered extensive academic attention, evidenced by a high citation count, hold historical significance in the field, mark a substantial advancement in technology, or be implemented in a significant real-world context. Recognizing the difficulty in evaluating the impact of newer AI systems since 2020 due to less available data, the authors also employ subjective judgment in their selection process for recent developments.
+
           - Systems are classified as "Industry" when their authors have ties to private sector entities, "Academia" when the authors come from universities or scholarly institutions, and "Industry - Academia Collaboration" if a minimum of 30% of the authors represent each sector.
         description_processing: >
           For each year starting from 1950, the total number of AI systems in each resercher affiliation category was calculated by adding that year's count to the previous years' counts. This provides a running total or cumulative count of AI systems for each year and researcher affiliation.

etl/steps/data/garden/artificial_intelligence/latest/epoch_aggregates_approach.meta.yml

@@ -23,6 +23,7 @@ tables:
         unit: 'AI systems'
         short_unit: ''
         description_key:
+            - The authors of the Epoch dataset have established a set of criteria to identify key AI systems, which they refer to as notable. These systems must demonstrate the ability to learn, show tangible experimental results, and contribute advancements that push the boundaries of existing AI technology. In terms of notability, the AI must have garnered extensive academic attention, evidenced by a high citation count, hold historical significance in the field, mark a substantial advancement in technology, or be implemented in a significant real-world context. Recognizing the difficulty in evaluating the impact of newer AI systems since 2020 due to less available data, the authors also employ subjective judgment in their selection process for recent developments.
             - Self-supervised learning is a machine learning technique where the model learns from the data itself without requiring external labels or annotations. It leverages inherent structures or relationships within the data to create meaningful representations. Self-supervised learning is commonly used in natural language processing and computer vision tasks, where models learn to understand context and semantics from large unlabeled datasets.
             - Unsupervised learning is a machine learning paradigm where the AI system explores patterns and structures within data without the presence of labeled examples. It aims to discover hidden relationships or groupings in the data. Unsupervised learning is applied in clustering, dimensionality reduction, and anomaly detection tasks. It's used when there are no predefined labels for the data.
             - Reinforcement learning is a type of machine learning where an agent learns to make decisions by interacting with an environment. It receives feedback in the form of rewards or penalties for its actions and aims to maximize the cumulative reward over time. Reinforcement learning is commonly used in robotics, game playing (e.g., AlphaGo), and autonomous systems where agents must learn to make sequential decisions.
@@ -37,6 +38,7 @@ tables:
         unit: 'AI systems'
         short_unit: ''
         description_key:
+            - The authors of the Epoch dataset have established a set of criteria to identify key AI systems, which they refer to as notable. These systems must demonstrate the ability to learn, show tangible experimental results, and contribute advancements that push the boundaries of existing AI technology. In terms of notability, the AI must have garnered extensive academic attention, evidenced by a high citation count, hold historical significance in the field, mark a substantial advancement in technology, or be implemented in a significant real-world context. Recognizing the difficulty in evaluating the impact of newer AI systems since 2020 due to less available data, the authors also employ subjective judgment in their selection process for recent developments.
             - Self-supervised learning is a machine learning technique where the model learns from the data itself without requiring external labels or annotations. It leverages inherent structures or relationships within the data to create meaningful representations. Self-supervised learning is commonly used in natural language processing and computer vision tasks, where models learn to understand context and semantics from large unlabeled datasets.
             - Unsupervised learning is a machine learning paradigm where the AI system explores patterns and structures within data without the presence of labeled examples. It aims to discover hidden relationships or groupings in the data. Unsupervised learning is applied in clustering, dimensionality reduction, and anomaly detection tasks. It's used when there are no predefined labels for the data.
             - Reinforcement learning is a type of machine learning where an agent learns to make decisions by interacting with an environment. It receives feedback in the form of rewards or penalties for its actions and aims to maximize the cumulative reward over time. Reinforcement learning is commonly used in robotics, game playing (e.g., AlphaGo), and autonomous systems where agents must learn to make sequential decisions.

etl/steps/data/garden/artificial_intelligence/latest/epoch_aggregates_domain.meta.yml

@@ -24,21 +24,23 @@ tables:
         short_unit: ''
         description_short: Refers to the specific area, application, or field in which an AI system is designed to operate.
         description_key:
-          - Games systems are specifically designed for games excel in understanding and strategizing gameplay. For instance, AlphaGo, developed by DeepMind, defeated the world champion in the game of Go. Such systems use complex algorithms to compete effectively, even against skilled human players.
+          - The authors of the Epoch dataset have established a set of criteria to identify key AI systems, which they refer to as notable. These systems must demonstrate the ability to learn, show tangible experimental results, and contribute advancements that push the boundaries of existing AI technology. In terms of notability, the AI must have garnered extensive academic attention, evidenced by a high citation count, hold historical significance in the field, mark a substantial advancement in technology, or be implemented in a significant real-world context. Recognizing the difficulty in evaluating the impact of newer AI systems since 2020 due to less available data, the authors also employ subjective judgment in their selection process for recent developments.
+
+          - Game systems are specifically designed for games and excel in understanding and strategizing gameplay. For instance, AlphaGo, developed by DeepMind, defeated the world champion in the game of Go. Such systems use complex algorithms to compete effectively, even against skilled human players.
 
           - Language systems are tailored to process language, focusing on understanding, translating, and interacting with human languages. Examples include chatbots, machine translation tools like Google Translate, and sentiment analysis algorithms that can detect emotions in text.
 
-          - Multimodal systems excel at processing multiple types of data at once, like sound and visuals together.  An example would be voice assistants like Siri or Alexa, which can understand spoken words (audio) and provide visual feedback (graphics) at the same time.
+          - Multimodal systems are artificial intelligence frameworks that integrate and interpret more than one type of data input, such as text, images, and audio. ChatGPT-4 is an example of a multimodal system, as it has the capability to process and generate responses based on both textual and visual inputs.
 
           - Vision systems focus on processing visual information, playing a pivotal role in image recognition and related areas. For example, Facebook's photo tagging system uses vision AI to identify faces.
 
-          - Speech systems are dedicated to handling spoken language, being the backbone of voice assistants and similar applications.  They recognize, interpret, and generate spoken language to interact with users.
+          - Speech systems are dedicated to handling spoken language, serving as the backbone of voice assistants and similar applications. They recognize, interpret, and generate spoken language to interact with users.
 
-          - Recommendation systems offer suggestions based on user preferences, seen prominently in online shopping and media streaming. For instance, Netflix's movie suggestions or Amazon's product recommendations are powered by such algorithms that analyze users' preferences and past behaviors.
+          - Recommendation systems offer suggestions based on user preferences, prominently seen in online shopping and media streaming. For instance, Netflix's movie suggestions or Amazon's product recommendations are powered by algorithms that analyze users' preferences and past behaviors.
 
-          - Drawing systems can create illustrations or sketches, either mimicking human techniques or generating unique art. Examples range from AI-generated artwork to design tools that can sketch based on user input or descriptions.
+          - Drawing systems can create illustrations or sketches, either by mimicking human techniques or by generating unique art. Examples range from AI-generated artwork to design tools that can sketch based on user input or descriptions.
 
-          - Other category represents a diverse set, including tasks from 3D reconstruction, Driving, Video, Text-to-Video, Search, Audio and Robotics.
+          - The 'Other' category represents a diverse set of tasks, including 3D reconstruction, autonomous driving, video processing, text-to-video synthesis, search algorithms, audio processing, and robotics.
         description_processing: >
           Processing involved calculating total number of AI systems in each domain for each year.
 
@@ -56,22 +58,23 @@ tables:
         short_unit: ''
         description_short: Refers to the specific area, application, or field in which an AI system is designed to operate.
         description_key:
-          - Games systems are specifically designed for games excel in understanding and strategizing gameplay. For instance, AlphaGo, developed by DeepMind, defeated the world champion in the game of Go. Such systems use complex algorithms to compete effectively, even against skilled human players.
+          - The authors of the Epoch dataset have established a set of criteria to identify key AI systems, which they refer to as notable. These systems must demonstrate the ability to learn, show tangible experimental results, and contribute advancements that push the boundaries of existing AI technology. In terms of notability, the AI must have garnered extensive academic attention, evidenced by a high citation count, hold historical significance in the field, mark a substantial advancement in technology, or be implemented in a significant real-world context. Recognizing the difficulty in evaluating the impact of newer AI systems since 2020 due to less available data, the authors also employ subjective judgment in their selection process for recent developments.
+
+          - Game systems are specifically designed for games and excel in understanding and strategizing gameplay. For instance, AlphaGo, developed by DeepMind, defeated the world champion in the game of Go. Such systems use complex algorithms to compete effectively, even against skilled human players.
 
           - Language systems are tailored to process language, focusing on understanding, translating, and interacting with human languages. Examples include chatbots, machine translation tools like Google Translate, and sentiment analysis algorithms that can detect emotions in text.
 
-          - Multimodal systems excel at processing multiple types of data at once, like sound and visuals together.  An example would be voice assistants like Siri or Alexa, which can understand spoken words (audio) and provide visual feedback (graphics) at the same time.
+          - Multimodal systems are artificial intelligence frameworks that integrate and interpret more than one type of data input, such as text, images, and audio. ChatGPT-4 is an example of a multimodal system, as it has the capability to process and generate responses based on both textual and visual inputs.
 
           - Vision systems focus on processing visual information, playing a pivotal role in image recognition and related areas. For example, Facebook's photo tagging system uses vision AI to identify faces.
 
-          - Speech systems are dedicated to handling spoken language, being the backbone of voice assistants and similar applications.  They recognize, interpret, and generate spoken language to interact with users.
-
-          - Recommendation systems offer suggestions based on user preferences, seen prominently in online shopping and media streaming. For instance, Netflix's movie suggestions or Amazon's product recommendations are powered by such algorithms that analyze users' preferences and past behaviors.
+          - Speech systems are dedicated to handling spoken language, serving as the backbone of voice assistants and similar applications. They recognize, interpret, and generate spoken language to interact with users.
 
-          - Drawing systems can create illustrations or sketches, either mimicking human techniques or generating unique art. Examples range from AI-generated artwork to design tools that can sketch based on user input or descriptions.
+          - Recommendation systems offer suggestions based on user preferences, prominently seen in online shopping and media streaming. For instance, Netflix's movie suggestions or Amazon's product recommendations are powered by algorithms that analyze users' preferences and past behaviors.
 
-          - Other category represents a diverse set, including tasks from 3D reconstruction, Driving, Video, Text-to-Video, Search, Audio and Robotics.
+          - Drawing systems can create illustrations or sketches, either by mimicking human techniques or by generating unique art. Examples range from AI-generated artwork to design tools that can sketch based on user input or descriptions.
 
+          - The 'Other' category represents a diverse set of tasks, including 3D reconstruction, autonomous driving, video processing, text-to-video synthesis, search algorithms, audio processing, and robotics.
         description_processing: >
           For each year starting from 1950, the total number of AI systems in each domain was calculated by adding that year's count to the previous years' counts. This provides a running total or cumulative count of AI systems for each year and domain.
 

snapshots/artificial_intelligence/latest/epoch.csv.dvc

@@ -19,15 +19,15 @@ meta:
       — Important state-of-the-art advance
       — Deployed in a notable context
 
-      The authors note that: "For new models (from 2020 onward) it is harder to assess these criteria, so we fall back to a subjective selection. We refer to models meeting our selection criteria as 'milestone models.
+      The authors note that: "For new models (from 2020 onward) it is harder to assess these criteria, so we fall back to a subjective selection. We refer to models meeting our selection criteria as 'milestone models."
     # Citation
     producer: Epoch
     citation_full: 'Epoch, ‘Parameter, Compute and Data Trends in Machine Learning’.
       Published online at epochai.org. Retrieved from: ‘https://epochai.org/mlinputs/visualization’
       [online resource]'
     # Files
     url_main: https://epochai.org/mlinputs/visualization
-    url_download: 
+    url_download:
       https://docs.google.com/spreadsheets/d/1AAIebjNsnJj_uKALHbXNfn3_YsT6sHXtCU0q7OIPuc4/export?gid=0&format=csv
     date_accessed: 2023-10-26
     # License