- Rating
- 1
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This paper introduces MOSAIC, a neurosymbolic model that represents each concept by multiple prototypes instead of a single one. The model defines "Implicit Reasoning Shortcuts" as predicting correct concepts and task labels but "mispredicting" the intended latent mode, and provides theory for recovering these modes.
Critical strengths
The paper presents an interesting combination of neurosymbolic AI and Infinite Mixture Prototypes, which allows the model to adaptively discover multiple prototypes for each concept.
Critical weaknesses
The central motivation of the paper is that each concept may require multiple modes in latent space and that existing prototypical NeSy approaches do not support this.
The two main critical weaknesses are that (1) the importance of recovering (and existence of) these modes is unclear and the theoretical motivation is not convincing, and (2) the characterization of prior work is misleading, since many existing models already use multiple prototypes per predicted variable or concept.
Regarding (1):
- The distinction between modes and ordinary variation around modes is unclear. Already in the running examples, it is unclear why different representations of writing a number (e.g. "3" vs "III") should be different modes, while different handwriting styles (e.g. "7" with or without a bar) should be variations near a single mode. Similarly, why pedestrians and bicycles could simply not be different variations of one mode for the concept "obstacle". This choice seems arbitrary. There is also no general expressivity argument, as a sufficiently expressive encoder can map all such variations near a single mode. If the different "desired modes" have independently justified human semantics, then the authors should explain why they should not instead be represented as distinct concepts (e.g. "bicycle", "pedestrian", "roman 3", "normal 3").
- The problem seems unrelated to reasoning shortcuts. The proposed definition of an Implicit Reasoning Shortcut (IRS) does not really define a reasoning shortcut. An IRS can occur even when the model predicts both the explicit concepts and the downstream task perfectly, even when there are no logic rules in the architecture, etc. Reasoning shortcuts are a form of misalignment that can happen because the supervision signal of the task combined with the logic may be insufficient to learn the right concepts.
- Additionally, Definition 2.1 is extremely brittle because it relies on exact equality of conditional distributions, which is very unlikely to hold in practice. There are many reasons why these distributions may differ that have nothing to do with reasoning shortcuts. For instance, any miscalibration of the model or approximation error would be classified as a "reasoning shortcut" according to this definition.
Regarding (2):
- The paper’s statements about prior prototype work is incomplete and misleading. PNet (Andolfi and Giunchiglia, 2025) indeed uses a single prototype per explicit concept, but many existing works do not have this limitation. In particular, the authors seem to claim that PGCM (Colamonaco et al., 2026) has the single-mode problem, while this is factually wrong, as PGCM seems to learn multiple prototypes per concept. PGCM even seems to define each concept explicitly as a disjunction over multiple prototypes. More generally, many prototype-based networks use multiple prototypes to predict a single variable, e.g. ProtoPNet (Chen et al., 2019). Therefore, the paper seems to primarily address a limitation from the PNet architecture rather than a general limitation of the literature, which significantly narrows the scope and novelty of the contribution.
Questions for authors
I refer to the weaknesses mentioned above.
15 comments
Comment (1) — Expressivity argument and definition of Implicit Reasoning Shortcuts.
3. On the importance of having such distinction in presence of a powerful encoder.
The reviewer claims:
“There is also no general expressivity argument, as a sufficiently expressive encoder can map all such variations near a single mode.”
We need to clarify that this is not true: this point has been addressed in Section 5.3. There we report a research question explicitly dedicated to the expressivity argument the reviewer refers to. The results already show that a deeper encoder can be beneficial for improving the concept F1-Macro score in (while remaining below MOSAIC’s), but it immediately backfires on . Indeed, a deeper encoder quickly overfits the limited support available and exhibits poor generalisation over the complex street scenes featured in the task. We further elaborate on this point below, and we provide additional results obtained with a deeper encoder (ResNet18) than the one reported already in the submitted manuscript (ResNet12). The settings are the same as Table 22 (, DPL baseline, support-set size 20 per concept).
| DPL | Acc(C) ↑ | F1(C) ↑ | Acc(Y) ↑ | F1(Y) ↑ | Cls(C) ↓ |
|---|---|---|---|---|---|
| +PNet | 72.52 ± 1.61 | 54.81 ± 0.93 | 58.41 ± 2.61 | 43.80 ± 1.79 | 0.00 ± 0.00 |
| +ResNet12-PNet | 50.16 ± 5.98 | 39.52 ± 4.95 | 47.40 ± 0.13 | 25.88 ± 0.25 | 0.07 ± 0.08 |
| +ResNet18-PNet | 45.87 ± 8.28 | 36.01 ± 6.15 | 47.38 ± 0.09 | 25.80 ± 0.09 | 0.13 ± 0.08 |
| +MOSAIC | 89.13 ± 0.12 | 57.59 ± 1.29 | 62.45 ± 3.02 | 41.42 ± 3.17 | 0.00 ± 0.00 |
As we can see, the more expressive encoder further degrades the performance of the PNet with the standard four-stack convolutional encoder and the shallower ResNet12 encoder. In particular, the mean concept collapse Cls(C) we obtained with the ResNet18 nearly doubles the one of ResNet12. Notice that from a theoretical perspective, by Corollary 4.7, no increase in capacity prevents an encoder that maps all modes onto the same prototype from falling into an IRS.
4. On the definition of Implicit Reasoning Shortcuts.
Our Definition 2.1 directly generalises the traditional definition of reasoning shortcuts [1] to the level of latent modes. Specifically, we distinguish between explicit reasoning shortcuts (ERSs), corresponding to traditional RSs where incorrect explicit-concept predictions nevertheless maximise the task likelihood, and implicit reasoning shortcuts (IRSs), where the predictor fails to recover the intended latent modes, up to permissible permutations. When each explicit concept has a single mode, our definition reduces to the traditional one. More generally, every ERS implies an IRS, but the converse does not hold.
The reviewer's observation that "an IRS can occur even when the model predicts both the explicit concepts and the downstream task perfectly" is correct and, indeed, precisely the additional phenomenon our generalisation is intended to capture.
The connection to traditional RSs lies in the semantic ambiguity left unresolved by the available supervision. In conventional RSs, the task supervision combined with the symbolic knowledge is insufficient to uniquely determine the intended explicit concepts. In our setting, this ambiguity extends to the underlying modes: the symbolic knowledge is defined over explicit concepts, and the downstream likelihood depends on the modes only through the taxonomy mapping them to explicit concepts. Consequently, mode assignments preserving the explicit concepts are indistinguishable to the downstream supervision. Thus, even when the task supervision is sufficient to recover the correct explicit concepts, it need not identify their underlying modes. Proposition 4.1 characterises the resulting deterministic optima, while Corollary 4.7 establishes that mode-blind prototypical predictors cannot recover the intended latent-mode structure, even under complete mode coverage.
Importantly, the relevance of IRSs goes beyond recovering latent structure. Although an IRS does not necessarily imply incorrect explicit-concept predictions, the mode-blindness underlying these failures can also compromise explicit-concept learning. By representing heterogeneous realisations through a single prototype, mode-blind predictors impose a restrictive inductive bias that can lead to incorrect explicit-concept assignments under limited supervision. Our existing experiments already demonstrate the practical relevance of this distinction. For example, on Omniglot-EvenOdd (10 alphabets), Table 3 shows that CCN+ with MOSAIC improves mode purity from 2.93% to 80.98% and reduces IRS-rate from 97.14% to 2.94%, while Table 1 shows a simultaneous improvement in explicit-concept F1 from 65.12% to 87.12%. Furthermore, Figure 5 shows that increasing support coverage allows MOSAIC to discover additional modes, accompanied by improvements in explicit-concept accuracy.
Comment (1) — Additional experimental analysis about the Implicit Reasoning Shortcuts definition.
The results we included in the answer above demonstrate that modelling the finer-grained structure is not merely an additional representational objective, but provides an effective inductive bias for improving the performance, as measured by accuracy and F1 score on the explicit concepts on which symbolic reasoning operates.
Additional empirical analysis.
To further illustrate why our generalisation is meaningful, we conducted additional experiments on measuring how frequently IRSs coexist with correct explicit-concept and downstream task predictions. Specifically, we introduce two metrics:
a. Concept accuracy under IRS, (: the proportion of IRS-affected images for which the explicit concept is correctly predicted.
b. Joint concept–task accuracy under IRS, (: the proportion of IRS-affected images for which both the explicit concept and downstream label are correctly predicted.
| Model | Mode purity (global) ↑ | IRS-rate ↓ | Query imgs in an IRS ↓ | Acc(C | IRS) ↑ | Acc(C,Y | IRS) ↑ |
|---|---|---|---|---|---|
| NN | 1.61 ± 0.31 | 98.57 ± 0.49 | 99.85 ± 0.19 | 27.29 ± 5.36 | 23.51 ± 4.66 |
| SL | 1.13 ± 0.07 | 99.14 ± 0.00 | 99.96 ± 0.13 | 8.47 ± 2.56 | 5.60 ± 2.34 |
| Scallop +C+E+R+PRE | 0.64 ± 0.35 | 99.60 ± 0.26 | 100.00 ± 0.00 | 15.79 ± 3.40 | 10.90 ± 1.55 |
| CTSketch +C+E+R+PRE | 0.58 ± 0.19 | 99.63 ± 0.13 | 100.00 ± 0.00 | 16.07 ± 3.98 | 10.61 ± 1.90 |
| CCN+ +C+E+R+PRE | 1.10 ± 0.24 | 99.23 ± 0.18 | 99.96 ± 0.13 | 28.76 ± 8.02 | 8.46 ± 6.41 |
| Scallop + PNet | 3.00 ± 0.06 | 97.14 ± 0.00 | 100.00 ± 0.00 | 69.90 ± 1.14 | 48.25 ± 1.72 |
| CTSketch + PNet | 3.03 ± 0.04 | 97.14 ± 0.00 | 100.00 ± 0.00 | 70.28 ± 1.08 | 48.63 ± 1.49 |
| CCN+ + PNet | 3.02 ± 0.05 | 97.14 ± 0.00 | 100.00 ± 0.00 | 69.21 ± 0.89 | 47.42 ± 1.27 |
| Scallop + MOSAIC | 77.56 ± 1.25 | 3.66 ± 1.37 | 7.09 ± 2.85 | 90.32 ± 3.11 | 79.56 ± 2.93 |
| CTSketch + MOSAIC | 77.72 ± 0.99 | 3.66 ± 0.74 | 6.92 ± 1.46 | 90.92 ± 4.12 | 80.20 ± 3.51 |
| CCN+ + MOSAIC | 77.69 ± 1.27 | 2.94 ± 0.51 | 5.70 ± 1.30 | 91.49 ± 3.39 | 81.18 ± 2.49 |
All values are percentages (mean ± standard deviation over 10 runs). These results substantiate the reviewer's observation: for CTSketch + PNet, 70.28% of IRS-affected images have the correct explicit concept and 48.63% have both the correct concept and downstream label. Such latent-mode errors remain invisible to conventional concept and task accuracy metrics, despite being captured by Definition 2.1. Moreover, mode-blind predictors map 350 underlying characters into at most 10 predicted categories, yielding an IRS-rate of at least (1-10/350=97.14%), consistent with the limitations characterised by Corollary 4.7. Conversely, MOSAIC substantially reduces this rate by explicitly representing multiple modes per concept.
We also agree that latent-mode misalignment can arise without a symbolic reasoning layer, as illustrated by the additional NN baseline, realising a knowledge-agnostic version of the Semantic Loss baseline (same architecture, cross-entropy loss over the downstream label rather than the semantic objective). Our claim is not that logical reasoning necessarily causes IRSs, but that the semantic ambiguity studied in traditional RSs extends to the latent structure underlying explicit concepts. Our theoretical analysis characterises this ambiguity, while our experiments demonstrate that explicitly modelling such structure can also improve the recovery of the explicit concepts required for reasoning.
We will revise the manuscript to clarify this connection and the distinction between ERSs and IRSs.
While we believe the terms implicit and explicit reasoning shortcut captures the generalisation of traditional RSs, we would welcome alternative terminology if the reviewer feels it would make both this distinction and their connection clearer.
Comment (1) — Modes relevance and their distinction from stylistic variations.
We thank the reviewer for their time and address the points raised below.
The content we present under each point will be included in the next version of the paper to improve its presentation.
We first answer the reviewer’s comment (1) in four points below.
Regarding comment (1)
1. On the importance of recovering the modes.
Learning latent fine-grained structure from coarse-grained supervision is a well-established problem in AI. Many learning problems involve an underlying hierarchy of concepts in which annotations are available only for coarse-grained categories, while the finer-grained categories remain unlabelled and must be inferred from data. This problem has been studied across several areas of AI. In inductive logic programming, target concepts can be represented as disjunctions of more specific sub-concepts [3], which can themselves be invented from examples of higher-level predicates [2]. Similarly, in computer vision, categories are often organised into coarse-to-fine hierarchies [4], while unannotated subpopulations within a class give rise to the well-studied problem of hidden stratification [5]. In natural language processing, learning fine-grained categories from coarse-grained supervision has also been explored as a means of reducing annotation costs [6]. MOSAIC addresses this problem in the neurosymbolic setting, where supervision is provided through symbolic knowledge defined over explicit (coarse-grained) concepts, while their underlying finer-grained realisations remain implicit. Rather than requiring these finer-grained categories to be separately annotated and introduced into the symbolic vocabulary, MOSAIC aims to automatically discover and represent them as latent modes with no additional supervision, while preserving the original explicit concepts used for reasoning.
2. On the distinction between modes and ordinary variations.
The distinction between modes and stylistic variations is formalised by our generative assumptions. Following [1], our mode-aware generative process (Figure 2) assumes that inputs are generated as x=ϕτ(m,s), where ϕτ is invertible and smooth with respect to the style variables s (Assumption [A2]). Smoothness implies that continuous changes in style produce continuous changes in the input. Consequently, for a fixed mode m, whenever the support of S is connected, the set of generated inputs Xm=ϕτ(m,s)∣s∈supp(S) is also connected. Moreover, invertibility ensures that distinct modes generate disjoint sets of inputs, i.e., Xm∩Xm′=∅ for m=m′.
This provides a principled interpretation of the distinction: stylistic variations describe continuous changes within a fine-grained category, whereas modes represent discrete factors of variation that cannot be explained by changes in style alone. For instance, variations in the stroke width or slant of an Omniglot character can be captured by s, whereas Arabic 3 and Roman III represent distinct fine-grained categories associated with the same explicit concept. Similarly, buses and construction vehicles constitute different possible realisations of the explicit concept obstacle in BDD-OIA. Observe these are example meant to provide an intuition to the reader. However, in the method we do not uniquely determine which fine-grained distinctions should be modelled as modes: this depends on the underlying generative factors and the chosen granularity of the concept hierarchy. Given such a decomposition, MOSAIC provides an inductive bias for preserving its heterogeneous structure through multiple prototypes, rather than requiring all realisations of an explicit concept to be represented by a single prototype. We will clarify this distinction and its underlying assumptions in the revised manuscript.