🔭 astrophysics

A Bayesian framework for posterior bounds on self-replicating probes in the Solar System under observational incompleteness

This paper presents a Bayesian framework that converts current observational null-detections and survey completeness data into quantitative, falsifiable upper bounds on the number of self-replicating probes in the Solar System, demonstrating that future constraints will be driven primarily by the discovery of sub-100m objects rather than kilometer-scale surveys.

Original authors: Jiawei Pan, Anqi Yang, Qiutao Gu, Yuan Zhang, Linglong Jiang, Jian Shi

Published 2026-07-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jiawei Pan, Anqi Yang, Qiutao Gu, Yuan Zhang, Linglong Jiang, Jian Shi

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the Solar System as a massive, messy attic filled with boxes, dust bunnies, and forgotten toys. For decades, scientists have been looking through this attic, hoping to find a specific type of toy: a self-replicating robot probe built by an alien civilization. So far, they haven't found a single one.

This paper is like a very careful accountant who takes that "nothing found" result and turns it into a mathematically strict rule about how many of those alien robots could be hiding there without us having seen them yet.

Here is the breakdown of their work using simple analogies:

1. The Big Question: "Are they hiding?"

The authors are tackling the Fermi Paradox. If advanced aliens exist and build robots to explore the galaxy, shouldn't we have seen at least one robot in our own backyard (the Solar System) by now? Since we haven't, either they don't exist, or they are very good at hiding.

2. The Method: The "Alien Robot Census"

Instead of just guessing, the authors built a Bayesian framework. Think of this as a super-precise calculator that combines two things:

The Guess (The Prior): Based on previous theories, they start with a wide range of possibilities for how many robots might exist. They assume the number could be anywhere from "almost zero" to "millions," but they weigh the possibilities carefully.
The Search (The Data): They looked at seven different "rooms" in our Solar System attic:
1. Near-Earth space (where asteroids come close to us).
2. Earth's co-orbital zones (objects sharing Earth's orbit).
3. Lagrange points (gravitational parking spots between Earth and Moon).
4. The Moon's surface.
5. The Main Asteroid Belt.
6. The Kuiper Belt (the icy outer rim).
7. The deep, distant outer Solar System.

They checked the "completeness" of the search in each room. For example, in the Main Asteroid Belt, we have looked at almost every large rock (over 1km), so the search is 99% complete. But in the deep outer solar system, we have barely looked at anything small.

3. The Calculation: "The Invisible Robot"

The authors used a formula to ask: "If there were 100 robots, how likely is it that our current telescopes would have missed all of them?"

The Size Matters: If the robots are huge (like a 1km skyscraper), we would have definitely seen them in the well-searched areas. So, the math says there are almost certainly zero of those.
The Small Robots: If the robots are tiny (like a 10-meter car), our telescopes might have missed them, especially in the dark, distant corners of the attic.

4. The Results: The "Upper Limit"

The paper doesn't say "Aliens are here" or "Aliens aren't here." Instead, it gives a maximum limit (a ceiling) on how many could be hiding.

For 10-meter robots: The math says there could be up to 14 of them hiding right now without us knowing. (It's unlikely there are 15, because we would have seen at least one).
For 100-meter robots: The limit drops to 3.
For 1-kilometer robots: The limit drops to less than 1. This means it is statistically impossible for a 1km robot to be hiding here without us seeing it.

The Catch: The result depends heavily on where the robots are hiding.

The "Lurker" Scenario: If aliens hide their robots in the hard-to-see places (like the deep outer solar system), the limit is higher (more could be hiding).
The "Diffuse" Scenario: If robots are spread evenly everywhere, the limit is lower (fewer could be hiding).

5. The Future: What Happens in the Next Decade?

The authors looked at upcoming space missions (like the Vera C. Rubin Observatory and the NEO Surveyor) which will act like brighter flashlights for the attic.

The Good News: These new missions will drastically improve our ability to find small robots (10 meters). The "ceiling" on how many could be hiding will drop by about 7.6 times over the next decade.
The Bad News: These missions won't change the limit for huge robots much, because we already know there aren't any huge ones hiding. We are already 99% sure of that.

The Main Takeaway

The most important part of this paper isn't the specific number "14." It's the tool they built.

They created a transparent, reproducible "calculator" that turns "we didn't find anything" into a strict scientific rule. They proved that to really solve the mystery of whether alien robots are here, we don't need to look for giant ships; we need to get much better at finding small, 10-meter objects in the deep, dark corners of our Solar System.

If we find even one 10-meter robot in the next ten years, the whole "upper limit" rule breaks, and we know for sure they are here. If we don't, we just know the number is even lower than before.

Technical Summary: A Bayesian Framework for Posterior Bounds on Self-Replicating Probes in the Solar System

Problem Statement
Despite decades of systematic searches, no robust evidence of extraterrestrial artefacts (self-replicating probes) within the Solar System has been reported. While the "Fermi paradox" has been extensively analyzed regarding the prevalence of civilizations in the Galaxy, there is a lack of a unified, end-to-end inference that quantitatively translates the absence of detections in specific Solar System regions into a falsifiable constraint on the expected number ( $\lambda$ ) of resident probes. Existing frameworks either focus on the Drake equation parameters for Galactic civilizations (e.g., Sandberg et al., 2018) or qualitatively discuss search targets without integrating heterogeneous survey completeness data into a single posterior distribution.

Methodology
The authors construct a Bayesian inference framework that combines a parametric prior with a Poisson likelihood derived from observational null results.

Prior Construction: The model adopts a log-uniform prior on $\lambda$ , derived from a product of Drake-equation-style factors (star formation rate, fraction of habitable planets, probability of intelligence, etc.) extended to include probe-specific parameters: the fraction of civilizations launching probes ( $f_{probe}$ ) and the steady-state number of probes per system ( $\langle n_{probe/sys} \rangle$ ). This results in a fat-tailed prior with a 95% upper limit of $\lambda_{prior} \approx 230$ .
Likelihood Function: The likelihood is modeled as a single-trial Poisson zero-detection event: $L(\lambda) \propto e^{-\lambda p_{det}}$ . Here, $p_{det}$ represents the effective probability that a single probe, if present, would have been detected by existing surveys.
Decomposition of Detection Probability ( $p_{det}$ ): The framework decomposes $p_{det}$ $p_{d e t}$ across seven distinct Solar System regions: Near-Earth Objects (NEOs), Earth co-orbitals, Earth–Moon Lagrange points, the lunar surface, the main asteroid belt, the Kuiper belt, and the outer/miscellaneous Solar System.
- For each region, a size-dependent completeness curve $p^{(i)}(d)$ is constructed by inverting or interpolating published survey efficiency data (e.g., Harris & Chodas for NEOs, Gregg & Wiegert for Lagrange points).
- A residence-weight scenario ( $w_i$ ) is applied to weight the contribution of each region to the total $p_{det}$ . Three scenarios are tested: "Default" (outer-Solar-System dominated), "Diffuse" (uniform volume distribution), and "Lurker" (concentrated in Earth co-orbitals).
Implementation: The inference is executed via Monte Carlo sampling ( $N=2 \times 10^5$ ) using a Python package (seta_bayesian). The code includes a rigorous test suite verifying prior bounds, completeness monotonicity, and convergence.

Key Results
The study reports 95% posterior upper bounds on $\lambda$ for three canonical probe sizes (10 m, 100 m, 1 km) under the three residence scenarios, based on data available as of 2025.

Current Constraints (2025):
- 10 m probes: Under the default residence model, the 95% upper bound is $\lambda < 14.3$ . The bound is loose because the effective detection probability is low ( $p_{det} \approx 0.0019$ ), driven by the assumption that most probes reside in the outer Solar System where survey completeness is negligible at this size.
- 100 m probes: The bound tightens to $\lambda < 3.16$ .
- 1 km probes: The bound tightens significantly to $\lambda < 0.68$ . At this scale, the posterior is "prior-saturated" because existing surveys (NEO, Main Belt) are already near-complete ( $>90\%$ ) for objects of this size; the data adds little information beyond the prior's upper tail.
Sensitivity: The "Lurker" scenario (probes in Earth co-orbitals) yields the tightest bounds (e.g., $\lambda < 2.27$ for 10 m), while the "Diffuse" scenario yields intermediate results. The "Default" scenario yields the loosest bounds due to the high weight assigned to poorly surveyed outer regions.
Future Forecasts (2025–2035): The framework projects the impact of upcoming missions (LSST, NEO Surveyor, Roman Space Telescope).
- For 10 m probes, the default-residence bound is forecast to compress by a factor of 7.6× (tightening the bound from ~14.3 to ~1.9) by 2035. This compression is highly sensitive to the realization of completeness targets below 100 m (envelope: 4.8×–10.0×).
- For 1 km probes, the bound moves by only ~14–16%, confirming that km-scale surveys are already in the prior-dominated regime.

Significance and Claims
The authors explicitly state that the primary contribution is not a definitive probability that probes exist, but rather a reproducible, data-driven framework that:

Quantifies Null Results: It converts heterogeneous, non-detection survey data into a rigorous posterior constraint on probe abundance.
Audits Assumptions: By using YAML-pinned priors and explicit provenance for every completeness input, the framework makes prior assumptions and data sources transparent and auditable.
Identifies the Critical Discovery Channel: The analysis reveals that the next decade's ability to constrain the existence of probes depends almost entirely on sub-100 m discovery (specifically the 10 m regime). Surveys targeting km-scale objects provide diminishing returns because the prior already saturates the bound at those sizes.
Defines the Regime Structure: The framework distinguishes between a "prior-saturated" regime (large objects, where data adds little) and a "data-constrained" regime (small objects, where improved survey completeness directly tightens the bound).

The paper concludes that while current data cannot rule out the existence of small probes, the framework provides a falsifiable upper bound that will tighten significantly if upcoming surveys achieve their stated completeness targets for objects smaller than 100 meters.