“Plasticity as the Mirror of Empowerment” (Abel et al., 2025) defines information-theoretic measures for plasticity (how much an agent is influence by its environment) and empowerment (how much the agent can influence the environment). Towards the end of the paper the authors make a claim that deterministic environments result in zero plasticity. This was initially very counterintuitive to me, so in this post I explore their definitions, show why their claim is true, and provide some intuition.


The authors start by defining agents as maps from observations to actions, and environments as maps from actions to observations . They define plasticity as the influence of observations on future actions, and empowerment as the influence of actions on future observations. Both quantities rely on mutual information. i.e., how much does knowing one quantity reduce your uncertainty about the other?

Towards the end of the paper, the authors claim that deterministic environments result in zero plasticity. This was initially very surprising to me. Even in the simplest RL problems, an agent adapts its internal state (and therefore its actions) based on observations. How could plasticity be zero? In the next section I verify this this claim by writing out the math. Then, I provide intuition: when examining an agent-environment interaction solely in terms of distributions over actions and observations, a deterministic environment implies that conditioning on observations does not change the probability of actions. Given previous actions, observations become redundant in the conditioning set because they are fixed functions of those actions. Therefore, action distributions are not influenced by observations, and plasticity is zero.

This counterintuitive result is partly a consequence of the choice of abstractions: their framing can be thought of as bird’s eye view of agents and environments, where one has knowledge of the counterfactuals for both, i.e., “what could have been.” Now let’s go into the math:

The Zero Plasticity Proof

After lots of preliminary definitions, they introduce generalized directed information (GDI): a measure of how two past elements in one sequence of random variables influence future elements in another sequence. This provides the foundation for understanding how actions affect observations (empowerment) and observations affect actions (plasticity). Concretely, for any two sequences of discrete random variables and defined over intervals and (where denotes the number of elements), the generalized directed information (GDI) is given by:

where the term inside the summation is the conditional mutual information, representing the reduction in uncertainty about the current outcome provided by the relevant window of , conditioned on the full history of both sequences. This notation assumes precedes and is a possible cause of .

Then, the plasticity of an agent relative to an environment over interval is defined as the GDI between observations and actions:

This brings us to the claim: every deterministic environment forces zero plasticity for any agent , i.e., . To show this, let’s start by computing how much information a set of observations provides about an action in a deterministic setting. Consider the mutual information between and a subset of observations , conditioned on the “past” history of observations and actions . We write this as:

Expanding this using the definition of conditional mutual information yields:

Here, is the entropy of a distribution. This equation asks: how does knowing change my uncertainty about ?

The key insight is that conditioning on the observation is redundant in a deterministic environment. The value of is a fixed consequence of the actions and the known past observations . Since the actions and past observations are already in the conditioning set of Term 2, adding the resulting observation provides no additional information about the agent’s uncertainty in . Therefore, Term 1 equals Term 2, and their difference is zero:

Now, recall that plasticity is the sum of these conditional mutual information terms (via GDI). Consequently:

Interpretation

This was quite counterintuitive at first. Consider the agent’s perspective: as it receives new information, it changes its internal state (e.g., its Q-function). Clearly it’s adapting, so it must be plastic, right? We must instead take a third-person perspective, viewing distributions over actions alongside a deterministic function mapping actions to observations. Plasticity is a function of how much an action distribution changes when conditioned on previous observations. Since the observations are deterministic functions of actions, they provide no unique information and the change in the distribution is zero.

Crucially, adaptation to new observations is still happening and visible in how the action distribution changes as the agent receives observations. However, in this specific theoretical frame, this change is only a function of the previous action distribution, because there is no observation uncertainty. Any change in the action distributions is purely a function of previous action distributions. Therefore, from a third-person probabilistic sense, the observations did not influence the actions. There is no plasticity (according to their definition).

This definition does not describe the physical capacity of an agent to change via new information, e.g., synaptic plasticity. Instead, it builds upon the authors’ (very general) abstractions of agents and environments. It allows for a probabilistic understanding of how observations and actions interact, without considering the agent’s internal machinery. Exploring where these probabilities “live,” and what they imply about causality and agent design is still unclear to me. Very cool read.


Blogpost 17/100

Bibliography

Abel, D., Bowling, M., Barreto, A., Dabney, W., Dong, S., Hansen, S. S., Harutyunyan, A., Khetarpal, K., Lyle, C., Pascanu, R., Piliouras, G., Precup, D., Richens, J., Rowland, M., Schaul, T., & Singh, S. (2025). Plasticity as the Mirror of Empowerment. The Thirty-Ninth Annual Conference on Neural Information Processing Systems. https://openreview.net/forum?id=eOZFqyE9Ok