| AI Model Training | Train deep neural network models, such as VGG 16-layer network and Inception-V4, using the combined dataset to Generalize across multiple cancer types. |
| Active Learning and Labeling | Utilize active learning techniques to prioritize data that offer high information gain and explore each data point chosen, injecting Domain Expertise directly into the training process to ensure accurate and reliable labels. |
| Adapt to Human Values | Algorithms must adapt to human values, as new metrics and methods are developed to evaluate their behavior against quantifiable human values. |
| Algorithmic Processing | The fully automated system employs advanced machine learning algorithms to process a large corpus of scientific papers. |
| Analyze data | The algorithms analyze the data to predict the energy consumption of each resource based on the behavior of Occupants. |
| Assigning scores | Assign a score to each generated phenotype image based on its similarity to the examples in the database |
| Automated Referral | Flag cases with high uncertainty, surpassing the defined threshold τ, for referral to medical professionals. |
| Automatic Application | Automatically apply the trained AI model to each cancer type without human adjustment, providing TIL prediction results. |
| Automatic Computation | The AI system utilizes algorithms to automatically compute explanations that suggest changes to the human's model. |
| Autonomous Learning | The collaboration aims to reduce the burden on the human operator and enhance the robot's autonomous learning capabilities. |
| Balancing User Commands | The algorithm balances the need to follow User commands closely while also deviating from the User's actions when they are suboptimal, discarding actions whose values fall below some threshold and selecting the remaining action closest to the User's input. |
| Calculate cosine similarity | Uses an embedding of the entire table to calculate the cosine similarity between the column names and semantic types. |
| Choose model | Choose one of the models among 'Search', 'Hand curated answers' etc. |
| Collect data | The algorithms collect data from IoT sensors and cyber-physical systems sensing/actuation platforms. |
| Comply with the Algorithmic Social Contract | Algorithms must comply with the algorithmic social contract that is programmed and monitored by Institutions and tools, ensuring that they operate in accordance with societal expectations. |
| Confidence level calculation | Calculate confidence levels for predicted image categories |
| Conversation Navigation | The dialogue system navigates the conversation to its successful completion using rule-based graph and learning elements |
| Convert Predictions for Display | The network's predictions are converted back to a format for display in WSI viewing software, such as Aperio ImageScope. |
| Create annotations | Create annotations for instance-aware semantic segmentation of reasonable quality with minimal effort. |
| Data Classification and Preprocessing | Classify MRI sequences using an ensemble of NLP and CNN models. Preprocess the MRI data, including image registration, skull stripping, and bias field correction. |
| Data Collection | Utilize input from Domain Experts to collect accurate annotations at scale |
| Define performance scores | Defining scores for overall, vehicle-trajectory, and handling performance based on the outcome of the optimizers. |
| Draw attention to region | The algorithm draws the human's attention to that region of the recording. |
| Evaluate maximum possible force of each tire | Evaluating the maximum possible force of each tire independently using the optTire optimizer. |
| Extract passages from documents | The system uses an AI model to extract passages from documents that may contain answers to the questions. |
| Extraction of Action Phrases and Initiation of Sentences from API Description | Algorithm extracts phrases from OpenAPI specification and possibly augments them with human-provided phrases. |
| Feedback Loop | As Users interact with the system and explore the retrieved scientific papers, their interactions and feedback can be used to further refine the algorithms and improve the relevance and accuracy of the results over time. |
| Filtering and Selection | Algorithm filters the generated sentences to eliminate noisy ones and selects a diverse set of sentences based on human feedback. |
| Finding visual similarity | Use CNN classifier to find visual similarity between generated phenotype images and database of examples |
| Fine-tune Based on Human Feedback | The network is fine-tuned based on human feedback, leading to improved accuracy and reduced burden of manual WSI annotation. |
| Fine-tune Language Model | Fine-tune the LM to theatre plays to see how far this approach can go. |
| Generate Final Text | Generate the final text using a similar approach as the base LM generation. |
| Generate candidate solutions | The EMO algorithm generates a set of candidate solutions. |
| Generate draft segmentation | The machine learning algorithm generates a draft segmentation of the video. |
| Generate new solutions | The EMO algorithm generates a new set of candidate solutions based on the guidance from the LTR model. |
| Generate new training data | Generates new training data using data programming and the inferred labeling functions. |
| Generate plan/COA | Automated planning system (RADAR) receives the model of the world, initial state, and goals/tasks from the human Decision maker. RADAR's automated planner analyzes the provided information and generates a plan or course-of-action (COA) based on the given model and goals. |
| Generate verbal explanations | Based on difficult-to-formalize experiential prior knowledge from human Domain Experts |
| Generate visual explanations | The system provides transparent explanations for the results to the human Domain Experts. |
| Generating curriculum | If no human input is possible, a different AI agent can be used to generate the curriculum. |
| Guide EMO algorithm | The LTR model is applied to guide the EMO algorithm towards the SOI. |
| Human-in-the-Loop Intervention | Incorporate human feedback to improve the model's performance and refine the uncertainty thresholding process. |
| Identify images with tortoises | The algorithm identifies at least one image containing tortoises per ground truth segment. |
| Identifying Differences | The AI system identifies differences between its own model and the human's model. |
| Image prediction | Predict categories of images using deep learning models trained from previous periods |
| Implicit Inference | The algorithm learns to assist the User without access to private information, implicitly inferring it from the User's input. |
| Improvement | CTN consistently improves with more human corrections, potentially achieving better performance than fully supervised methods with considerably less annotation efforts. |
| Incorporate Societal Values | Algorithms must incorporate the values and goals negotiated by humans and Stakeholders, considering trade-offs between different societal interests and ethical considerations. |
| Incorporating Inferred Goals | The algorithm is capable of incorporating inferred goals into the agent's observations when the goal space and User model are known, further improving sample efficiency. |
| Incorporation | CTN formats these corrections as partial contours and incorporates them back into the training via an additional Chamfer loss. |
| Infer labeling functions | Infers labeling functions from data used to generate new training data and prediction functions. |
| Inference | Once we have elicited samples from the Expert prior, we can use deep learning techniques to infer the correct label for a given input x. This involves training a deep model on the labeled data and using it to predict labels for new inputs. |
| Initialization | CTN takes the exemplar contour as an initialization and gradually evolves it to minimize the weighted loss for each unlabeled image. |
| Integrate feedback | From the human Domain Experts to improve its performance |
| Integrating Labeled Behaviors | The algorithm integrates the labeled behaviors provided by Players into the computational models of esports Players, incorporating qualitative and quantitative inputs. |
| Intent Classification | The NLU system classifies User queries into intent categories using natural language understanding techniques |
| Issue Identification and Response Delivery | The chatbot system identifies challenges in providing a good response and may hand off the conversation to a human agent or raise a ticket for human intervention, aiming to provide consistent and suitable responses while delivering a high-quality customer experience |
| Iterative Improvement | The curiosity agent continues to select actions and seek out new information, iteratively improving the robot's learning process. |
| Learn preference | The LTR neural network learns the Decision Maker's preference based on the feedback. |
| Learning | The algorithm uses deep reinforcement learning with neural network function approximation to learn an end-to-end mapping from observation and input to agent action values, with task reward as the only form of supervision. |
| Learning from human demonstrations | Agents learn from human demonstrations under the imitation learning (IL) paradigm when it is challenging to design a reward function or when the reward function could be sparse, thus making it hard for an RL agent to learn. |
| Learning from human-generated tasks | Agents learn from the tasks generated by humans. |
| Machine Learning-driven Interpretation | Utilizes deep learning models to interpret the natural language input and understand the User's instructions. |
| Make predictions on new data | Uses the local model to make predictions on new data. |
| Match column names to type ontology | Matches each column name to the labels in the type ontology using syntactic and semantic matching. |
| Model Deployment | Trained intent classifier is deployed and serves to classify Users’ intents based on their input utterances in the dialog system while attempting to invoke an API endpoint conversationally. |
| Model Training | Algorithm trains the intent recognition model for the chatbot. |
| Model updates | Include pseudo-labels in final data set for further model updates or ecological analyses |
| Monitor planning process | The automated planner continuously monitors the planning process of the human Decision maker and the current state of the environment. |
| Natural Language Generation | Algorithm generates equivalent sentences using a variety of language models. |
| Operate Transparently | Algorithms must operate transparently, allowing for debugging and monitoring of their behavior to ensure that they align with the algorithmic social contract. |
| Optimize energy consumption | The algorithms optimize the energy consumption of the building as a whole. |
| Optimize vehicle acceleration in Cog | Optimizing the vehicle's acceleration in the center of gravity (Cog) using the optCog optimizer. |
| Optimizing Plans and Behavior | With the reconciled model, the AI system optimizes its plans and behavior to be in line with the updated human model. |
| Outcome Assessment and Reporting | Provide segmentation masks for longitudinal tumor tracking and quantitative growth assessment. Generate standardized reports and visualizations based on the processed MRI data. |
| Performance Benchmarking | Provide performance benchmarks to encourage the development of accurate and interpretable downstream models for the computational analysis of H&E stained colon tissue |
| Present candidate solutions | The consultation module presents a set of selected candidate solutions to the Decision Maker. |
| Present new candidate solutions | The consultation module presents the new set of candidate solutions to the Decision Maker. |
| Presenting Explanations | The AI system presents these explanations to the human, highlighting the suggested changes to the human's model. |
| Private Experimentation and Collaboration | Run AstronomicAL entirely locally on the User’s system, providing a private space to experiment. Export a simple configuration file to share entire layouts, models, and assigned labels with the community, allowing for complete transparency and effortless reproduction of results. |
| Process Whole Slide Images | The algorithm processes whole slide images using a semantic segmentation network. |
| Processing Observed Behaviors | The algorithm processes observed behaviors and generates data representations of Player actions within the game environment. |
| Producing phenotype images | Generate a set of phenotype images |
| Propose machine learning model architecture | Proposing a machine learning model architecture to directly determine the scores from the initial data. |
| Provide alerts/suggestions | Based on the current state and resource availability, RADAR's automated planner provides alerts and suggestions to the human Decision maker regarding potential drawbacks in the plan, resource constraints, or problems that may arise in the future. |
| Provide answer | Provide the answer which is observed by the CSA (care Agent). |
| Provide feedback | Once we have inferred the correct label for a given input x, we can provide feedback to the student based on their performance. This feedback can be associated with specific parts of a student's solution and can articulate their misconceptions in the language of the instructor. |
| Provide feedback | The algorithms provide feedback to Occupants about their energy usage. |
| Provide interpretations | The interpretations provided by the computer should help the human better decide whether to trust the computer's prediction or not. |
| Providing Feedback | The algorithm provides real-time action feedback to the User based on the learned mapping. |
| Purpose Matching | Algorithms are utilized to emulate human judgment of purpose match, ensuring that the system finds partial purpose matches in the top results. |
| Ranking phenotype images | Rank the generated phenotype images in order of aesthetic quality based on the assigned scores |
| Re-training on artist-specific datasets | Increase accuracy in automating personal aesthetic judgement |
| Reformulate NLP Task | Tackling the generation of clarifying questions for truly interactive agents. |
| Request Ground Truth | The curiosity agent may request ground truth annotations from the human operator when additional information is needed. |
| Restrict Generation | Restrict the generation by enforcing that only certain predetermined characters speak, possibly in a pregenerated order. This can be achieved by stopping the generation. |
| Retrain with Corrected Annotations | The corrected annotations provided by human Experts are used to retrain the semantic segmentation network. |
| Reward Function Decomposition | The algorithm decomposes the agent's reward function into known terms computed for every state and a terminal reward provided by the User, enabling the system to learn efficiently from a dense reward signal that captures Generally useful behaviors and adapt to individual Users through feedback. |
| Route questions to appropriate respondents | The system routes questions to the appropriate respondents, either the AI model or human Knowledge workers. |
| Seed Language Model | Seed a language model (LM) with a prompt that is the beginning of a dramatic situation. |
| Segment object from background | Automatically segment the corresponding object from the background using basic image processing techniques. |
| Select Actions | The curiosity agent selects actions to navigate the robot within the exploration space. |
| Semi-Automatic Annotations | Apply a classification algorithm to Whole Slide Images (WSIs) and adjust the predicted TIL probability maps by applying thresholds to generate semi-automatic annotations. |
| Simplify Data Collection | Using a single-turn data collection strategy to increase the speed of data collection. |
| Sketch Generation | Generates sketched scenes based on the interpreted text instructions, using state-of-the-art deep neural network architectures. |
| Speed Up Training Environment | Using a new gridworld environment for fast and scalable experiments. |
| Suggest ways to improve behavior | The algorithms suggest ways to improve the energy-saving behavior of Occupants. |
| Test Set Creation and Validation | Curate a labeled test set to demonstrate the validity and Generalizability of the model. Mark any example as unsure, ensuring that all training data are of high quality. |
| Train local model | Trains a local model on the new training data. |
| Training Data Generation | Combine manually annotated patches and semi-automatically annotated patches to form the training set for the AI model. |
| Training on database of examples | Learn visual features important for aesthetic evaluation |
| Tumor Segmentation and Feature Extraction | Segment tumor tissue subtypes using CNNs to generate quantitative tumor measurements. Optionally allow Expert-in-the-loop manual refinement of segmentation results. |
| Uncertainty Estimation | Generate stochastic predictions using Monte Carlo Dropout, capturing aleatoric and epistemic uncertainties in the segmentation outputs. |
| Uncertainty Thresholding | Apply an uncertainty threshold (τ) to the model's predictions, determining the level of uncertainty for each case. |
| Update guesses | The computer periodically updates its guesses and interpretations (every 4 words in the experiments described in the PDF). |
| Update the model | The model is updated based on the feedback provided by the Experts, and the learning process continues iteratively. |
| Update weights | Update its weights based on this feedback. |
| User Feedback and Refinement | Incorporates User feedback and iteratively refines the sketches based on the provided instructions. |
| User Interface | The system presents the top results to the Users through an intuitive and User-friendly search interface. |
| Utilize Feedback | The robot uses the feedback to improve its performance and adjust its exploration strategy. |
| Utilizing Corrected Model | The algorithm utilizes the corrected computational model to refine the understanding of Player intents, strategies, and tactics within the gaming environment. |
| Validation of Annotations | Compute quantitative concordance statistics between pathologists and the dataset to ensure the accuracy of the annotations |