iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Pedro Domingos’s five tribes of machine learning are Symbolists, Connectionists, Evolutionaries, Bayesians, and Analogizers. They are intellectual traditions—not companies, teams, or exclusive categories—and describe different ways machines can learn from experience. There is no universally best tribe; the right approach depends on the problem, data, uncertainty, explainability needs, and operating constraints.
What are the five tribes of machine learning?
Domingos presents the five traditions in The Master Algorithm. Each emphasizes a different way to represent knowledge or improve a model. The labels are a useful mental map, not a formal taxonomy: methods can overlap, and real systems can combine ideas from several tribes.
| Tribe | How it learns | Representative methods | Useful comparison |
|---|---|---|---|
| Symbolists | Learn explicit rules, concepts, and structured relationships. | Decision trees, random forests, production rules, inductive logic programming, and knowledge graphs. | How inspectable the model’s reasoning is. |
| Connectionists | Adjust weights in networks inspired by connections between neurons. | Artificial neural networks, deep learning, transformers, and some reinforcement-learning approaches. | Pattern-recognition capability versus explainability. |
| Evolutionaries | Search by generating variations, evaluating them, and iterating through selection. | Genetic algorithms, evolutionary programming, genetic programming, and evolutionary strategies. | Optimization and search across possible designs or solutions. |
| Bayesians | Represent uncertainty with probabilities and update beliefs as evidence arrives. | Bayesian networks, probabilistic models, hidden Markov models, and some approaches to causal inference. | Uncertainty handling and use of prior knowledge. |
| Analogizers | Use similarity to known examples to make predictions or decisions. | k-nearest neighbors, support-vector machines, case-based reasoning, and recommendation methods. | Similarity, retrieval, and example-based classification. |
These examples are representative, not rigid assignments. A method’s place in the map depends on the idea being emphasized, and a deployed system may use several approaches together. The overview of the tribes and their methods is discussed by BMC and TechBloat.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsWhat does each tribe do?
Symbolists: rules and structured knowledge
Symbolist approaches work with explicit concepts and relationships. A decision tree, for example, can express a sequence of conditions leading to a result. This can make a model’s logic easier to inspect, although interpretability varies by method and by the size and complexity of the resulting model.
#1 Best Overall
- Use scikit-learn to track an example ML project end to end
- Explore several models, including support vector machines, decision trees, random forests, and ensemble methods
- Exploit unsupervised learning techniques such as dimensionality reduction, clustering, and anomaly detection
- Dive into neural net architectures, including convolutional nets, recurrent nets, generative adversarial networks, autoencoders, diffusion models, and transformers
- Use TensorFlow and Keras to build and train neural nets for computer vision, natural language processing, generative models, and deep reinforcement learning
Connectionists: learned representations
Connectionist methods learn patterns by adjusting the weights in networks. Deep learning is mainly part of this tribe because it uses layered neural networks. Neural networks can recognize complex patterns, but their decisions may be difficult to explain; BMC characterizes this as the “proverbial black box.”
Evolutionaries: search by variation and selection
Evolutionary methods repeatedly generate and evaluate candidate solutions, keeping or modifying candidates according to how well they meet an objective. They are useful when the task is framed as searching a large design or optimization space.
Rank #2
Bayesians: reasoning with uncertainty
Bayesian approaches express uncertainty probabilistically and revise estimates as evidence changes. They can incorporate prior knowledge, making them useful when uncertainty itself matters to a decision.
Analogizers: learning from similar examples
Analogizer methods compare a new case with known cases and use those similarities to classify, recommend, or otherwise make a prediction. Their behavior depends in part on how similarity is defined and which examples are available.
Which tribe is best for a particular problem?
There is no universal winner. Asked whether interpretable models should always beat black-box approaches, Domingos answered, “It depends on the application.” On how to select a tribe, he noted that “no one has a good theoretical answer to this problem,” while pointing to practical heuristics and trying alternatives where appropriate. These comments appeared in a 2015 KDnuggets Q&A.
Start by comparing the approaches against the requirements of the task rather than choosing a tribe by reputation:
Rank #4
- Representation: Does the problem naturally involve explicit rules, patterns in large amounts of data, a search objective, probabilities, or similarity to prior cases?
- Interpretability: Do people need to inspect or explain how a prediction was reached?
- Uncertainty: Must the system express how uncertain it is, or incorporate prior knowledge?
- Data and compute: What training examples and computational resources are available?
- Optimization behavior: Is the central challenge finding a good solution among many possible designs?
- Change over time: Will the data or operating environment shift, and how will the system be maintained?
- Error costs and constraints: What happens when the system is wrong, and what operational or policy requirements must it meet?
These considerations help narrow the choices, but they do not guarantee a single best method. When practical constraints allow, compare plausible alternatives on the actual task and evaluate them against the consequences of their errors.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallWhich tribe does deep learning belong to?
Deep learning belongs mainly to the Connectionist tribe: it uses layered neural networks whose weights are adjusted from data. That classification identifies its central learning idea, not everything that might be involved in a modern AI system. Deep-learning systems can be paired with symbolic rules, probabilistic reasoning, similarity-based retrieval, or optimization methods.
Best Value
Do modern AI systems combine ideas from multiple tribes?
Yes. The tribes are not exclusive camps, and a production system can use different approaches for different parts of a task. BMC illustrates this with a self-driving system that could use connectionist methods for sensor perception, evolutionary methods to search for a driving policy, analogizer methods to account for driver types, and symbolist rules to represent road constraints.
This is why the framework remains useful as a mental map: it helps distinguish the kinds of learning and reasoning a system uses, even when its components do not fit neatly into one category. The map does not imply that every system uses all five or that one combination is best for every application.
Where can you learn more?
Domingos’s book, The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World, is the central source for his five-tribe framework. KDnuggets identifies the book as the source in which he explains the tribes.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
