INDUCTIVE BIAS:
A bias commonly described in studies that use machine learning, but also relevant in human psychology. The set of assumptions or rules, that an algorithm (or learner) uses to predict outcomes, when it is given a set of inputs (i.e. data/information) it has not previously seen1,2,3. In other words, one can say: a learner is biased to his/her/its assumptions.
Inductive Bias is not avoidable, or a choice of the learner during decision making, and thus always relied upon. However, Inductive Bias can be changed prior to the decision making process. Inductive Bias in the context of machine learning is similar to Implicit Bias in the context of human psychology.
Inductive Bias (i.e. Implicit Bias) is manufactured in machine learning algorithms through the process of model development. Implicit Bias in humans is generated through life experiences. Also see: Implicit Bias, Automation Bias, Bias Cascade, Algorithmic Bias, Estimation Bias, and Forecast Bias.
References:
1. Bordelon B, Pehlevan C. Population codes enable learning from few examples by shaping inductive bias. Elife. 2022 16;11:e78606. (Link to Reference)
2. Török B, Nagy DG, Kiss M, Janacsek K, Németh D, Orbán G. Tracking the contribution of inductive bias to individualised internal models. PLoS Comput Biol. 2022;18(6):e1010182. (Link to Reference)
3. Sinz FH, Pitkow X, Reimer J, Bethge M, Tolias AS. Engineering a Less Artificial Intelligence. Neuron. 2019;103(6):967-79. (Link to Reference)