Deterministic Programming and Inductive Learning

38 min
Block 0 — Situating supervised learning
Objective
master the paradigm reversal that machine learning rests on, decide whether a given problem calls for a deterministic or an inductive approach, and justify that decision by the cost of errors and by the division of labor between human and machine.
Estimated duration
35 minutes
Prerequisites
chapter 001
Associated quizzes
002.1-quiz-paradigm-reversal.md to 002.7-quiz-human-machine-roles.md

1. The paradigm reversal

Chapter 001 established that machine learning is an approach to artificial intelligence founded on induction. This chapter works out what that formula means when you sit down to design a system.

1.1 The two schemas

Both paradigms handle the same three objects — rules, data, answers. They differ only in where each object sits in the production chain.

ParadigmInputs suppliedOutput producedAuthor of the rules
Deterministic programmingRules + DataAnswersA human designer
Inductive learningData + AnswersRulesThe algorithm, from the observations

The discriminating point : in the first paradigm the rule is an input to the system; in the second it is the output.

That single displacement drives everything else in this chapter. It determines who is accountable for the rule, how the rule is verified, what happens when the world changes, and what kind of failure the system exhibits when it goes wrong.

1.2 What a decision rule is

Before comparing the two paradigms further, the object they both produce needs a definition that does not depend on how it was obtained.

DEFINITION — Decision rule

Rigorous definition

A deterministic mapping from a description space X to a decision space Y, which assigns to every description x in X exactly one decision in Y: formally, a function h : X → Y. In classification, Y is a finite set of categories; in regression, an interval of the real line.

In plain terms

A procedure that, for any situation you can describe, returns one decision and only one.

The essential point

The definition says nothing about the origin of the rule. An IF ... THEN ... clause, a statutory rate schedule, a tree learned on 200,000 case files and a neural network are all decision rules in this sense. They differ only in how h was obtained.

Point of caution

A model is no less a rule than a hand-written clause: it is a rule whose statement nobody ever wrote down. That property is the source of both its power and its resistance to audit.

The consequence is worth stating plainly. Choosing between the two paradigms is never a choice between "having a rule" and "not having a rule". Both approaches end with a rule in production. The choice is about who writes it, what evidence justifies it, and what it costs to keep it current.

1.3 Deduction and induction

The two paradigms rest on two distinct modes of inference, identified by logic long before computing existed.

DEFINITION — Deduction and induction

Deduction

The mode of inference in which a conclusion is derived from premises by transformation rules that preserve truth. If the premises are true and the derivation is valid, the conclusion is necessarily true. Illustration : "Every message carrying an executable attachment is blocked. This message carries one. Therefore it is blocked."

Induction

The mode of inference in which a general proposition is formed from a finite set of particular observations. The conclusion exceeds the content of the premises: it is not guaranteed, only more or less probable given what was observed. Illustration : "Across 40,000 observed messages, those combining certain terms were reported as unwanted in 96% of cases. Therefore a new message combining those terms is probably unwanted."

The problem of induction

Induction offers no logical guarantee. Statistical learning theory (Vapnik and Chervonenkis, 1970s) does not dispose of that objection: it bounds it with probabilistic guarantees, conditional on explicit assumptions, the principal one being that future data are drawn from the same distribution as the training data.

Operational consequence

A deterministic system is wrong when its rule is wrong. An inductive system is wrong some fraction of the time by construction, including when every step has been carried out correctly.

CriterionDeductionInduction
Direction of inferenceGeneral to particularParticular to general
Status of the conclusionNecessary if the premises are trueProbable, revisable
Content of the conclusionContained in the premisesExceeds the premises
Effect of a new observationNone on the ruleMay revise the rule
Corresponding paradigmDeterministic programmingMachine learning
Failure modeRule badly specifiedInsufficient generalization

The last row deserves emphasis. The two failure modes call for different engineering responses. A badly specified rule is found by reading the specification and reproducing the case; it is fixed once, and the fix is permanent. Insufficient generalization is found only by measurement on data the model never saw, and the fix — more data, different features, a different model family — is itself subject to measurement.

1.4 The unknown target function

Induction needs a formal object to aim at. Supervised learning postulates one.

DEFINITION — Target function f and hypothesis f-hat

Rigorous definition

We postulate the existence of an unknown target function f : X → Y linking the description space X to the space of values to be predicted Y. It is accessible only through a finite sample of observed pairs (x₁, y₁), ..., (xₙ, yₙ). The algorithm selects, from a hypothesis space H fixed by the choice of model family, a hypothesis f-hat that minimizes a measure of error on the sample, in the hope that f-hat approximates f across the whole domain.

In plain terms

There is a real relationship between what you observe and what you want to predict. Nobody knows its statement. You build an imitation of it from the examples you have.

The three gaps between f-hat and f

  • H may contain no function close to f — approximation error, tied to the choice of model.
  • The sample is finite and noisy — estimation error, tied to the data.
  • The relationship may change over time — drift, covered in chapter 082.

Point of caution

The existence of f is a working assumption, not an established fact. When the target depends on factors absent from the descriptions x, no function from X to Y is exact, whatever the algorithm. Section 5 develops this point.

Read the diagram from the top left. Only two boxes are ever observed: the sample and f-hat. The target function f is never seen, never printed, never validated directly. Everything a practitioner does — splitting data, measuring error, comparing models — is an indirect attempt to bound the distance between the one object the algorithm produced and the one object nobody can inspect.

This asymmetry explains a habit that outsiders find excessive: the insistence on evaluating a model on data withheld from training. It is not procedural formality. It is the only available evidence about a quantity that cannot be measured head-on.


2. Case study: how a rule-based system degrades

The following case reconstructs a typical trajectory in email filtering between the mid-1990s and the mid-2000s. The orders of magnitude are representative.

2.1 Chronology

Month 1 — three rules are enough.

IF subject contains "viagra"     THEN spam
IF subject contains "lottery"    THEN spam
IF body contains "click here"    THEN spam

Three rules, written in half a day, intercept 82% of unwanted messages at zero maintenance cost. The system is a success by every measure available at the time, and the decision to write rules rather than collect data was the correct one.

Month 3 — obfuscation. Senders start altering the spelling of the filtered terms: V1AGRA, V-I-A-G-R-A, VÍAGRA, \/IAGRA, Vi@gra, or the same letterforms in homoglyphic Cyrillic characters. The interception rate falls to 51% in three weeks. The team adds one rule per observed spelling and the corpus grows from 3 to 47 rules.

Month 6 — false positives. The accumulated rules start blocking legitimate mail: correspondence from a partner pharmacy, a medical newsletter followed by 4,000 customers, internal marketing campaigns that contain the words "click here". Three incidents reach senior management in a single month. The team no longer dares delete rules, since each one intercepts real spam. Instead it adds exceptions: sender allowlists and per-domain waivers. The corpus reaches 210 rules, 60 of which are exceptions to other rules.

Month 12 — new vectors. The text moves into an image, then into an attachment, then gets fragmented by invisible characters. Each vector demands its own family of rules: 610 rules.

Month 24 — unmanageability.

IndicatorMonth 1Month 6Month 12Month 24
Number of rules32106101,800
Interception rate82%79%77%76%
False positive rate0.1%0.9%1.6%2.3%
Staff assigned to maintenance0123
Response time to a new campaign1 day3 days7 days11 days
Share of rules with an identifiable author100%70%35%12%

Nobody on the team knows the full set of 1,800 rules, and any modification produces unpredictable regressions elsewhere in the corpus. Performance has fallen despite a threefold increase in maintenance effort.

Read the last row before the others. The share of rules whose author can still be identified drops from 100% to 12%. At that point the corpus has stopped being a specification and become an artifact: it encodes decisions nobody can reconstruct or justify, and it is therefore no longer safe to change.

2.2 The three degradation mechanisms

The failure cannot be blamed on the team's competence. It is structural.

MechanismEffect on the rule corpusGoverning quantity
Combinatorial explosionThe number of cases to cover grows faster than the capacity to write themNumber of variants of a single pattern
Coupling between rulesEvery new rule alters the behavior of existing rulesNumber of rule pairs
Adversarial obsolescenceRules expire on a schedule set by a hostile third partyGap between attack cycle and correction cycle

Mechanism 1 — combinatorial explosion. A single six-character term admits a number of equivalent spellings that grows multiplicatively. Count the plausible substitutions position by position for viagra: four for v (v, V, \/, and the Greek nu), four for i (i, 1, l, í), four for each of the two occurrences of a (a, @, á, 4), two for g (g, 9). That already gives 4 × 4 × 4 × 4 × 2 = 512 forms. Optionally inserting a separator at each of the five internal positions multiplies that count by 2⁵ = 32, for a total of 16,384 forms — for one word, before considering case variation or homoglyphs drawn from other alphabets.

Manual writing progresses linearly while the space to cover grows exponentially. No amount of maintenance effort closes that gap. This is the decisive asymmetry, and it does not depend on how good the team is.

Mechanism 2 — coupling between rules. A corpus of n rules contains n(n−1)/2 pairs that can interact.

Number of rules3472106101,800
Number of pairs31,08121,945185,7451,619,100

Confirming that a new rule conflicts with none of the others means examining 1,800 interactions on every addition. That check is not feasible by hand, so it is not performed. The regressions are discovered in production, by users.

Note the shape of the growth. Between month 12 and month 24 the corpus roughly triples, but the number of interacting pairs grows by a factor of nearly nine. The verification burden does not track the size of the system; it tracks the square of it.

Mechanism 3 — obsolescence against an adaptive adversary.

DEFINITION — Adaptive adversary

Rigorous definition

An agent whose objective function includes circumventing the decision system, and who receives an observation signal on that system's decisions allowing it to adjust its behavior.

In plain terms

Someone who benefits from fooling the system, who can see what gets through and what gets blocked, and who adapts accordingly.

The critical property: asymmetry of cycles

The sender tests a variant and observes the outcome within hours, at near-zero cost. The defending team must notice the decline, diagnose the vector, write the rule, verify that it breaks nothing, and deploy it — several days. As long as the attacker iterates faster than the defender, performance converges toward that of a system filtering only obsolete campaigns.

Point of caution

This property is not specific to rule-based systems: a model facing an adaptive adversary degrades too. The difference lies in the unit cost of an update — retraining is tooled and repeatable, rewriting 1,800 rules is not.

The three mechanisms compound. Combinatorial explosion sets how many rules must exist; coupling sets how expensive each one is to add safely; adversarial obsolescence sets how quickly the whole corpus expires. The month-24 numbers are the arithmetic consequence, not an accident of execution.

2.3 The inductive reformulation

Rather than specifying the patterns characteristic of spam, you assemble a corpus of messages whose nature is known and let the algorithm identify the discriminating regularities.

The loop from user reports back into the corpus is the structural difference. In the rule-based system, every new spam campaign is a cost. In the inductive system, every new spam campaign that a user reports is an input.

DimensionRule-based systemInductive system
Response to a new spellingWrite a dedicated ruleAbsorbed if the corpus is refreshed
Adaptation effortProportional to the number of variantsConstant: retrain
Source of the signalThe analyst who observesReports from users
Effect of traffic volumeRising maintenance loadAdditional resource for the model
Legibility of a decisionHigh: the triggered rule is namedReduced: a weighting of many indicators

Point of caution : the inductive reformulation does not eliminate the work. It moves the work to maintaining a labeled corpus, a retraining pipeline and performance monitoring. The gain comes from the fact that this work is tooled and repeatable, not from its disappearance.

That caution matters at the point of proposing the change. A team that pitches machine learning as a way to stop maintaining the filter has mis-sold the project. The correct pitch is that the maintenance becomes a fixed, automatable cost instead of one that grows with the adversary's imagination.


3. Choosing between the deterministic and the inductive approach

3.1 Decision tree

The order of the questions is not arbitrary. The first question is never "which algorithm should we use" but "is the rule known". Machine learning becomes relevant only when the answer is no.

Three of the four terminal states in the tree are not machine learning. That proportion is closer to professional reality than the volume of published material on the subject would suggest.

3.2 The six conditions of applicability

ConditionJustificationIllustration
The rule is not formalizableThe decision criteria cannot be put into words, not even by the experts who apply themRecognizing a face, judging a composite risk
Factors and interactions are numerousManual writing diverges and coupling between rules becomes unmanageableCredit scoring on 60 variables
The environment changesWritten rules expire; retraining absorbs the change at constant costFraud detection, email filtering
Personalization must operate at scaleOne rule per individual is impossible to write and to maintainRecommendation across several million accounts
An approximate answer is acceptableThe model produces an estimate carrying uncertainty, never a certaintyPrice estimation, demand forecasting
Sufficient labeled history existsInduction requires observations whose target value is knownTransactions with an observed outcome

These six conditions are conjunctive. A problem that satisfies five and fails the sixth — typically the absence of labeled history — is not a machine learning problem as it stands.

The word "as it stands" carries the practical content. Failing the sixth condition is often a temporary state that instrumentation can fix within two or three quarters. Failing the first — the rule is perfectly formalizable — is permanent, and no amount of data collection changes the verdict.

3.3 The five contraindications

ContraindicationReasonAlternative
The exact rule is known, stable and verifiableThe model would substitute an approximation for an exact answer, with nothing gained in returnImplement the formula
Exactness is a legal or contractual obligationA statistical approximation is legally inadmissible on an amount owedExplicit rule, audited, versioned
The volume of observations is insufficientInduction on too small a sample produces a rule that does not generalizeDomain expertise, prior data collection
The error is catastrophic and unrecoverableNo model reaches 100%: the residual fraction of errors is structuralSystematic human review
No history exists because the process is newThere is nothing to induce fromPrototyping, provisional rules, instrumented collection

3.4 The three situations where the deterministic approach wins

DEFINITION — Auditability of an automated decision

Rigorous definition

The property of a decision system that allows, for any decision produced, the reconstruction of the chain of justification that led to it and of the exact state of the system on that date, reproducibly by a third party. It requires that the applied rule be statable, that the version in force be retained, and that the result be reproducible identically.

In plain terms

Being able to show, months later and to someone outside the team, exactly why this decision came out this way and to obtain the same result again.

Point of caution

A statistical model satisfies reproducibility but makes the statement of the rule difficult: post hoc explanation methods produce an approximate justification of the decision, not its statement. When the obligation bears on the individual motivation of a decision, that approximation may be insufficient.

Situation 1 — the rule is known, exact and stable. Computing a sales tax, converting at a published rate, the installment schedule of a fixed-rate loan, validating an identifier against its check digit. A model trained to reproduce these operations would reach 99.7% accuracy where direct implementation reaches 100%. The substitution is a pure regression.

Situation 2 — no history is available. A process created last week has no observations whose target value is known. The question is not which algorithm to choose but how to instrument the process so that observations get collected and labeled.

Situation 3 — auditability is an obligation. When a decision must be individually justified to the person affected or to a regulator, the ability to state the applied rule becomes a design constraint on the same footing as performance. A model remains relevant as decision support, with the enforceable decision produced by an explicit rule or by an operator.


4. The cost of errors as the decision criterion

This section is the most important in the chapter. It separates a technical understanding from an operational one.

4.1 The base premise

No machine learning model reaches 100% accuracy on new data.

The statement follows from the inductive nature of the inference: the rule is extrapolated from a finite sample onto a domain that exceeds it. In practice, an announced accuracy of 100% on new data signals either data leakage, covered in chapter 028, or an evaluation carried out on the training data. The relevant question is therefore never "does the model make mistakes" but:

Who handles the errors, at what cost, and within what delay?

4.2 The calculation to run every time

A model reaching 95% accuracy is generally presented as a good model. Set against the volume processed:

Daily decision volume   : 100,000
Model accuracy          : 95%
Correct decisions       :  95,000
Erroneous decisions     :   5,000  per day
Handling assumptionLoad inducedConclusion
No handling, error without consequenceNoneModel deployable as is
Handled by the end user, 30 seconds each42 hours of attention per day, diffuseAcceptable if the error is visible and correctable
Internal human review, 4 minutes per case333 hours per day, roughly 42 full-time positionsThe recovery mechanism costs more than the model
Error not detectable by the userZero load, unbounded riskThe most dangerous configuration

The last row is the most important. A costly but visible error is a sizing problem. A silent error is a design problem.

The arithmetic is worth doing out loud, because the gap between rows two and three is where most deployment decisions actually turn. Thirty seconds of a user's attention on an error they can see and dismiss is an annoyance. Four minutes of an internal operator's time on the same volume is a department. The model is identical in both rows; only the recovery path differs.

The first branch of this tree comes before any question of volume or cost. If the error is not detectable, no amount of accuracy makes the system safe, because nothing in the system will ever report that it is failing.

4.3 Reading grid by context

ContextNature of the errorSeverityReversibilityML appropriate
Content recommendationIrrelevant suggestionLowImmediate: the user ignores itYes, without reservation
Email filteringFalse positive: legitimate mail quarantinedMediumGood if the quarantine is browsableYes, provided the quarantine is accessible
Predictive maintenanceSpurious alert, or a failure not anticipatedMedium to highPartial: cost of intervention or cost of downtimeYes, with an asymmetric threshold
Credit decisionsUnjustified refusal to a solvent applicantHighLow: harm suffered, litigation possibleYes, subject to explainability and an appeal path
Automated moderationRemoval of lawful contentHighConditional on an appeal path existingYes, with guaranteed human re-review
Medical diagnosisFalse negative: undetected pathologyVery highLow to noneNo, not as an autonomous decision: decision support
Payroll or tax computationIncorrect amountHighCorrectable, but a breach of an exactness obligationNo: the exact rule exists

Three readings of this table need to be mastered.

First reading : severity depends on the context of use, not on the error rate. The same 95% model is excellent for recommendation and unacceptable for payroll.

Second reading : reversibility is the decisive operational criterion. A severe but recoverable error is managed by design — a cautious threshold, a verification queue, an appeal path. An irreversible error cannot be managed that way.

Third reading : the two error types of a classifier do not carry the same cost. That asymmetry is a design parameter, adjustable through the decision threshold, and is developed in chapter 062.

ANALOGY — Quality control at the end of the line

A production line does not pursue zero defects at any price. It sizes an acceptable defect rate, then designs the detection and rework capacity to match. That rate is not a property of the machine: it follows from the unit cost of a defect reaching the customer, the cost of inspection, and the volume produced. A manager who demands zero defects without sizing the inspection does not get zero defects. He gets undetected defects.

The acceptable error rate of a model is, in the same way, an engineering decision — never a property of the algorithm.


5. The structural limits of machine learning

The four limits below are not implementation defects. They follow from the inductive nature of the approach, and no algorithm, no architecture and no volume of data removes them.

LIMIT 1 — The model creates no information that is absent from the data

Statement

A learning algorithm extracts regularities present in the descriptions supplied. It cannot base a decision on a factor that appears in no variable, either directly or through correlation.

Illustration

A property valuation model fed only floor area and bedroom count cannot price the view, the orientation, or the condition of the roof, which nonetheless determine a substantial share of the price. The model does not report the deficiency: it produces an estimate whose residual error will be attributed to "noise" when it in fact reflects missing information.

Practical consequence

The performance ceiling is set by the information content of the variables, not by the choice of algorithm. This limit explains a frequent disappointment: moving from a regression to a sophisticated ensemble method buys a few points at most when the necessary information is absent from the dataset.

LIMIT 2 — The model establishes no causality

Statement

A predictive model identifies statistical associations between explanatory variables and the target variable. An association is not a cause-and-effect relationship: a high-performing model may rest on associations whose causal interpretation is wrong, or even reversed.

The documented asthma and pneumonia case

A study from the 1990s on patients admitted for pneumonia set out to predict mortality risk, so that low-risk patients could be directed toward outpatient care. The model learned a rule that was exact on the data and dangerous in practice:

Pneumonia patients with a history of asthma have a lower risk of death.

The association is real in the data. Its explanation: those patients were systematically admitted to intensive care because of their history, received immediate and aggressive treatment there, and therefore survived at a higher rate. The induced rule was a consequence of the care protocol, not a clinical property. Applied as written, it would have sent home precisely the patients whom the existing protocol was protecting. The model was not faulty — it reproduced the data faithfully — and only a review by clinicians surfaced the anomaly.

The case is reported notably by Cooper and coauthors (1997), then taken up and analyzed by Caruana and coauthors (2015) in their work on intelligible models in healthcare.

The formulation to retain

A model predicts what happens under the conditions in which the data were produced. It does not predict what would happen if those conditions were acted upon. Any intervention decision based on a purely predictive model falls outside that model's domain of validity.

LIMIT 3 — The model assumes the relationship is stationary

Statement

The statistical framework of supervised learning assumes that production data are drawn from the same distribution as the training data. When that assumption stops holding, the performance guarantees fall with it.

Two forms of breakdown

  • Data drift : the distribution of the explanatory variables changes while the relationship between variables and target remains valid. Example: the customer base gets younger, and the observed profiles differ from those in the training corpus.
  • Concept drift : the relationship itself changes. Example: a purchasing behavior that used to signal loyalty becomes, after a change in the product lineup, a signal of imminent churn.

Practical consequence

Usual causes: changes in customer behavior, modification of the product or the business process, regulatory change, adaptation by an adversary, replacement of a sensor, external shock. A model in production therefore does not have constant performance: it has a shelf life, which must be monitored rather than assumed. Putting a model into production without a monitoring mechanism is an incomplete operational decision. Drift, its detection and retraining strategies are covered in chapter 082.

LIMIT 4 — Garbage in, garbage out, aggravated by silent failure

Statement

The quality of the induced rule is bounded by the quality and representativeness of the observations supplied. Data that are erroneous, biased or mislabeled produce a rule that is erroneous, biased or badly calibrated.

The aggravation specific to machine learning

A deterministic program fed invalid data fails visibly: an exception, an aberrant value, a rejected check. A learning algorithm does not fail. It converges, it produces a model, and that model produces predictions that look normal. The failure is silent.

Data defectEffect on the modelHow it shows up
Partially erroneous labelsThe induced rule reproduces the labeling errorApparent performance fine, decisions wrong
Unrepresentative populationThe rule holds only on a subgroupPerformance gap between segments
Variable unavailable at decision timeExcellent performance in test, collapse in productionData leakage, chapter 028
Strongly imbalanced classesThe model favors the majority classHigh accuracy, zero detection on the rare class, chapter 050

Point of caution

The absence of a runtime error is never evidence of validity for an inductive system. Validation requires inspecting the data, analyzing errors by segment, and confronting the learned regularities with domain knowledge.

Taken together, the four limits describe what a supervised model is and is not. It is a compressed record of the regularities present in one finite sample, described by one chosen set of variables, produced under one set of conditions, at one moment in time. Every limit above is a restatement of one of those four qualifiers.


6. A worked comparison: house price estimation

The objective is to estimate the sale price of a house. Both approaches are carried out side by side on the same properties.

6.1 The deterministic version

A domain expert proposes a formula grounded in experience:

price = area_m2 x 3,500 + bedrooms x 10,000
PropertyAreaBedroomsComputed priceActual priceErrorRelative error
A90 m²3$345,000$352,000−$7,0002.0%
B120 m²4$460,000$448,000+$12,0002.7%
C75 m²2$282,500$291,000−$8,5002.9%
D140 m²4$530,000$690,000−$160,00023.2%

On the first three properties the formula is acceptable. On the fourth it misses by $160,000. A gap of that size disqualifies the tool: no professional bases a negotiation on an estimate whose possible error reaches a quarter of the value.

Note what the first three rows would have done to a team that stopped there. Three consecutive errors under 3% look like a validated formula. The failure appears only on the property that differs structurally from the others, and the formula gives no indication that it is out of its depth on that property.

6.2 Diagnosing property D

FactorSituation of property DEffect on price
NeighborhoodSought-after school catchment, transit station 300 meters awayStrong, multiplicative
RenovationKitchen and bathroom redone 18 months before the saleAdditive, substantial
GarageDouble garage, uncommon in the areaAdditive, moderate

The correction looks simple: add the missing terms.

price = area_m2 x A + bedrooms x B + garage x C
      + recently_renovated x D + neighborhood_index x E + F

The structure is plausible. The problem lies elsewhere:

What are the values of A, B, C, D, E and F?

Nobody knows. They appear in no reference table, and they vary by city, by market segment and by period. An expert can propose an order of magnitude for A. He cannot adjudicate between $34,000 and $41,000 for the value of a renovation, nor quantify a neighborhood index on a scale coherent with the other terms. On top of that come interactions — the value of a renovation depends on the neighborhood — plus non-linearities, and instability: the coefficients will be out of date in eighteen months.

Conclusion : the structure of the problem is identifiable by an expert; its parameters are not. That is exactly the configuration that calls for an inductive approach.

This distinction between structure and parameters recurs throughout the course. Domain expertise is nearly always right about which variables matter and how they enter the model. It is nearly always unable to supply the numbers.

6.3 The inductive version

You assemble the completed sales, described by the same variables, with the observed price.

AreaBedroomsGarageRenovatedNeighborhood indexObserved price
90 m²3003$352,000
120 m²4103$448,000
75 m²2003$291,000
140 m²4117$690,000
105 m²3105$468,000
160 m²5116$712,000
82 m²2014$371,000

The real dataset holds several thousand rows of this shape, and the learning procedure fits in three statements:

python
X = sales[["area_m2", "bedrooms", "garage", "renovated", "neighborhood_index"]]
y = sales["price"]

model.fit(X, y)

The call model.fit(X, y) is the act of induction: the algorithm searches for the coefficients that minimize the gap between computed prices and observed prices across all the sales.

TermEstimated coefficientBusiness reading
Area$3,120 per m²Marginal value of a square meter
Bedrooms$6,400 per bedroomOwn effect, holding area constant
Garage$14,900Premium for a garage
Recent renovation$38,200Premium for a recent renovation
Neighborhood index$21,500 per pointEffect of the area
Intercept$18,700Value at the origin

None of these values was supplied. They were induced from the observed sales. The area coefficient differs from the expert's intuition — 3,120 rather than 3,500 — because the initial estimate was implicitly absorbing effects that the other terms now capture.

Prediction on a property never seen:

Property E : 110 m², 3 bedrooms, garage, not renovated, neighborhood index 4

110 x 3,120 = 343,200
  3 x 6,400 =  19,200
     garage =  14,900
 renovation =       0
 4 x 21,500 =  86,000
  intercept =  18,700
-----------------------
  Estimate  = $482,000

In professional use, this point estimate is reported together with an interval and an average error, both established during evaluation. An estimate delivered without either is not a professional deliverable, whatever the model behind it.

6.4 Comparing the two chains

CriterionDeterministic approachInductive approach
Origin of the coefficientsPostulated by an expert, unverifiableEstimated from observed sales
Taking a new factor into accountRewrite the formula, set the coefficient by handAdd a column and retrain
Adapting to a different marketRedo the expert elicitation entirelyRetrain on local data
Measuring qualityGap noticed after the fact, with no protocolError measured on sales not used for learning
Auditability of a decisionHigh: the formula is statableMedium to low, depending on the model family

Reading the table : the inductive approach wins on four criteria and loses on auditability. That trade-off is representative, and it is settled by the context of use. An indicative estimate for a real estate agent tolerates medium auditability. An appraisal value enforceable in court does not.

Note also the fourth row, which is easy to skim past. The deterministic formula has no protocol for measuring its own quality. Its errors surface one property at a time, in the field, with no aggregate anyone ever computes. The inductive approach does not merely produce better coefficients: it comes with a measurement discipline that the hand-written formula never had.


7. The division of labor between human and machine

Press vocabulary — "the algorithm decides", "the machine learns by itself" — sustains the idea that the decision chain is automated end to end. Examining the actual cycle of a project contradicts it.

7.1 Who decides what

Stage of the cycleContent of the decisionDecider
1. Frame the business problemDetermine which question is worth askingHuman
2. Rule on whether ML is warrantedTest the problem against the conditions in section 3Human
3. Define the target variableChoose what gets predicted and how it is measuredHuman
4. Choose and collect the dataDetermine sources, period, populationHuman
5. Define the performance measure and thresholdFix what counts as success, arbitrate between error typesHuman
6. Prepare the data and build the variablesClean, encode, aggregate, deriveHuman, tooled
7. Choose the algorithm familyTrade off performance, interpretability, cost, constraintsHuman
8. Fit the model parametersEstimate the coefficients minimizing error on the sampleMachine
9. Evaluate and interpret the resultsAnalyze errors, confront domain knowledge, detect anomaliesHuman
10. Decide on production deploymentRule on release and its safeguardsHuman
11. Monitor and arbitrate retrainingDetect drift, decide to restart the cycleHuman

Out of eleven stages, the machine covers one, and that one is entirely conditioned by the human decisions that precede it: the data from stage 4, the target from stage 3, the error measure from stage 5 and the model family from stage 7 fully determine what stage 8 is able to produce.

This is not a rhetorical deflation of machine learning. Stage 8 does something no human can do: it searches a parameter space of arbitrary dimension for the configuration minimizing a stated criterion, across millions of observations, in seconds. The point is that the value of what it produces is bounded by the ten decisions surrounding it.

7.2 The cycle and its decision zones

The dark blocks are human decisions. The light block is the only stage executed by the machine.

ANALOGY — The instrument and the experimenter

A spectrometer produces measurements no operator could obtain by eye. It is irreplaceable on that segment of the chain. It does not choose the research question, does not draw the sample, does not define the protocol, does not interpret the results and does not decide to publish. A learning algorithm occupies exactly the same position.

DEFINITION — Degree of autonomy of a decision system

Rigorous definition

The position of an automated system on the scale running from the mere production of information to a decision executed without human intervention, characterized by the role assigned to the operator inside the decision loop.

In plain terms

How much of the decision the system actually makes, and what the person is there to do.

The three usual regimes

RegimeRole of the humanContext of use
Decision supportThe human decides, the model informsSevere errors, motivated decisions, indispensable expertise
Human in the loop on exceptionThe model decides, the human arbitrates uncertain casesHigh volume, recoverable errors, identifiable uncertainty
Full automationThe model decides alone, the human monitors aggregatesLow-severity errors, massive volume, immediate reversibility

Point of caution

The autonomy regime is a design decision, independent of the model's performance. A very strong model may be deployed as decision support because the context requires it. A mediocre model may be fully automated because the error carries no consequence there.


8. Common reasoning mistakes

MISTAKE — Presenting machine learning as superior to classical programming

Symptom : framing the choice as a modernization, or contrasting an "outdated" method with a "current" one.

Why it is wrong : the two approaches answer two distinct configurations. Machine learning contributes nothing when the rule is known, exact and stable; it introduces an approximation there with nothing gained in return.

Correct formulation : "Deterministic programming and inductive learning address two different configurations. The criterion for choosing is the availability of an exact and stable rule, not the age of the method."

MISTAKE — Believing the model discovers the rules without human intervention

Symptom : describing the project as "we hand the data to the algorithm and it figures everything out".

Why it is wrong : the algorithm executes one stage out of eleven. The choice of target, of data, of performance measure and of model family is human, and it fully conditions the result obtained.

Correct formulation : "The algorithm estimates the parameters. Framing the problem, choosing the data, defining the performance measure and validating the result remain human decisions."

MISTAKE — Treating a learned association as a causal relationship

Symptom : deciding to act on a variable because the model assigns it a large weight — "the model shows that a history of asthma lowers the risk, so those patients can go home".

Why it is wrong : a predictive model reproduces the regularities of the process that generated the data, protocols and policies included. It says nothing about what would happen if that process were changed.

Correct formulation : "The model establishes an association valid under the conditions in which the data were produced. Any decision to intervene on a variable requires a separate causal analysis, with a protocol designed for it."

MISTAKE — Treating a high accuracy figure as proof of fitness

Symptom : concluding in favor of deployment on the strength of a single headline number, without examining the volume or the nature of the remaining errors.

Why it is wrong : 95% accuracy on 100,000 daily decisions is 5,000 errors per day. Whether deployment is warranted depends on their severity, their detectability, their reversibility and the cost of handling them.

Correct formulation : "The accuracy figure is not sufficient. You must establish the expected error volume, the severity of each error type, who detects them, and at what cost they are put right."

MISTAKE — Assuming a deployed model keeps its performance

Symptom : shipping a model with no monitoring, treating the project as finished at release.

Why it is wrong : the guarantees of supervised learning assume that production data follow the same distribution as the training data. That assumption stops holding as soon as the users, the product, the regulation or an adversary changes.

Correct formulation : "A model has a shelf life. Going to production includes performance monitoring and a retraining procedure." The subject is covered in chapter 082.


9. Summary

THE REVERSAL
    Deterministic programming : Rules + Data    -> Answers
    Inductive learning        : Data + Answers  -> RULES
    The rule is an INPUT in the first case, an OUTPUT in the second.

MODE OF INFERENCE
    Deterministic : deduction, conclusion necessary
    Inductive     : induction, conclusion probable and revisable

FORMAL FRAME
    f     target function, unknown, never observed
    f^    hypothesis selected by the algorithm, an approximation of f
    Gaps  : approximation (model), estimation (data), drift (time)

CRITERIA FOR CHOOSING, IN ORDER
    1. Is the exact rule known and stable?       -> deterministic
    2. Is there sufficient labeled history?      -> if not, collect it
    3. Is an approximate answer acceptable?      -> if not, deterministic
    4. Is the cost of errors controllable?       -> if not, decision support
    5. Otherwise                                 -> inductive approach

THE CALCULATION TO RUN EVERY TIME
    95% accuracy on 100,000 decisions = 5,000 errors per day
    Question : who handles them, at what cost, within what delay?
    Most dangerous case : the error that cannot be detected.

THE FOUR STRUCTURAL LIMITS
    1. No information is created if it is absent from the data
    2. A learned association is not an established cause
    3. Stationarity of the relationship is an assumption, not a fact
    4. Defective data -> defective model, with no error signal at all

DIVISION OF LABOR
    11 stages in the cycle. The machine covers 1:
    fitting the parameters. The other 10 are human decisions.

Summary statement

Deterministic programming and inductive learning are distinguished by the position of the decision rule: specified as an input in the first case, induced as an output in the second. The inductive approach is warranted only when the exact rule is out of reach, a labeled history exists, an approximate answer is acceptable, and the cost of the residual errors is controlled; within it the machine takes charge of one stage only, fitting the parameters, every other stage being a human decision.


Quizzes and interview questions : 002.1-quiz-paradigm-reversal.md to 002.7-quiz-human-machine-roles.md

Next chapter : 003.0-three-learning-paradigms.md