Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper argues that scientific hypothesis generation, specifically identity abduction, can occur without continuous sensorimotor embodiment by leveraging representational grounding. The authors introduce the Abduction Loop architecture, which facilitates the generation of hypotheses through representation generation, motif extraction, and cross-domain retrieval, demonstrating its effectiveness with a multimodal model that linked gravitational memory transport to weak-lensing cosmology. The proposed DAB-30 benchmark provides a falsifiable evaluation framework to assess the architecture's capabilities in generating and verifying scientific hypotheses.
Identity abduction can thrive without physical interaction, as shown by a model linking disparate scientific concepts through representational grounding.
Can scientific abduction occur without continuous sensorimotor embodiment? Recent arguments in AI and philosophy of science hold that genuine hypothesis generation requires an agent continuously coupled to the physical world. We defend a narrower claim: online embodiment is not necessary for every abductive scientific act. Our focus is identity abduction: the inference that two independently developed structures are one object under an explicit correspondence, reached through representational grounding rather than bodily interaction. An agent may acquire new inferential affordances not through physical interaction but through transformations into representations that expose latent invariants. Scientific diagrams are a practical substrate because they embody independently evolved conventions that partially canonicalize symmetry, topology, and operator structure across disciplines - a property we develop as convention space, which answers a hard retrieval problem: finding mathematically related work when two fields share no discriminating vocabulary. We operationalize the mechanism as an architecture, the Abduction Loop: representation generation, motif extraction, convention-space canonicalization, cross-domain retrieval, identity-hypothesis generation, and adversarial verification, with abstention as the designed default. A documented episode, in which a multimodal model given a figure of a gravitational-memory transport model generated and then verified the hypothesis that its central differential complex is equivalent to the spherical Kaiser-Squires mass-mapping complex of weak-lensing cosmology, serves as a motivating possibility witness from which the architecture is abstracted, not as evidence of general capability. We close with a falsifiable evaluation program, the DAB-30 benchmark. The contribution is a mechanistic proposal, an architecture, and a test program.