{"id":"7c197b84-3b2a-4777-9653-353a4268e89d","arxiv_id":"2504.17268","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A certified algebraic solver based on rational univariate representations outperforms homotopy continuation on some parameter estimation problems for ODE models, while homotopy continuation remains better on systems with many solutions.","lead":"This paper compares two ways of solving the polynomial equations that arise when estimating unknown parameters in ODE models from data. It finds that an exact algebraic solver can handle some models that a popular numerical solver cannot finish within a day, and vice versa.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Abstract's 'vice versa' is unsupported by Table 1: no example shows HC solving where RUR fails, and the comparison rests on one untested parameter set per model.","rationale":"The reader's verdict of CONDITIONAL is appropriate, but the most load-bearing concern is slightly different from the reader's stated weakest assumption. The reader emphasizes the single parameter set per model, which is indeed a serious generalization risk. I agree with that concern partially, but the more immediate issue is an internal overstatement: the abstract's 'vice versa' claim is contradicted by the paper's own table. No reported row has HC solving where RUR fails; the closest cases, Crauste-2 and Crauste-3, are solved by RUR albeit more slowly and with larger error. This means the central comparative claim as written is not fully supported by the evidence, independent of whether more parameter sets are tested. The single-parameter-set issue compounds this by making even the supported direction (RUR beating HC on Akt-2) potentially non-generalizable. Both concerns point to the same remedy: soften the abstract's binary 'out of reach' language, report multiple parameter sets, and clarify that the conclusions are instance-specific benchmarks rather than universal solver capabilities. The paper does provide reproducible code and a useful empirical comparison, so conditional acceptance with required revisions is the right outcome; no change to the reader's verdict is needed. I mark agreement as partial because the reader identified a real weakness but not the more directly falsifiable one that the 'vice versa' claim lacks any supporting row in Table 1.","tokens_in":3601,"tokens_out":4919,"duration_ms":44912,"concrete_test":"Rerun the benchmark for Akt-2 and Crauste-3 (and ideally all models) on 10 randomly sampled parameter vectors per model, using the same time and memory limits as the paper. Then check whether any parameter vector yields a case where HomotopyContinuation.jl returns a solution within the limits while the RUR solver does not; if no such case exists, the abstract's 'vice versa' clause is unsupported and should be removed or revised to 'sometimes faster and more accurate on certain examples.'","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim in the abstract is that the RUR solver can tackle examples out of reach for homotopy continuation 'and vice versa.' The paper's own Table 1 does not provide a single row where HC Julia succeeds (within the stated time/memory limits) while RUR fails. In the three examples where HC is faster and more accurate (Crauste-1, Crauste-2, Crauste-3), RUR still solves all of them; the paper itself labels Crauste-2 and Crauste-3 as 'More challenging for RUR,' not 'out of reach.' Thus the literal 'vice versa' direction of the central claim is not established by the reported experiments. The only supporting examples for HC superiority are artificially constructed by squaring parameters in Crauste-1, which the paper acknowledges: 'we had difficulties to illustrate that with our small set of examples. Thus, we construct Crauste-2 and Crauste-3 artificially.' Furthermore, all classifications of 'out of reach' depend on arbitrary resource limits (timeout > 1 day, memory > 100 GB) and on exactly one parameter set per model, with the untested assumption that 'the solvers behave similarly for different numerical values.' Since the polynomial systems are data-dependent, solver performance can vary substantially with parameter values, so the benchmark's conclusions may not generalize beyond the tested instances. The core comparative claim is therefore weaker and more conditional than the abstract implies.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript compares a new certified polynomial system solver based on rational univariate representation (RUR) and real root isolation with HomotopyContinuation.jl (HC) on parameter estimation problems in rational ODE models. The authors use a differential-algebra construction that turns the estimation task into a polynomial system, and benchmark both solvers on a set of models from the literature. Table 1 reports running times and parameter estimation errors. The main claim is that the RUR solver can solve systems that homotopy continuation cannot, and vice versa, with the RUR solver also being more reliable on some structurally non-identifiable models.","tokens_in":3873,"tokens_out":5471,"duration_ms":48396,"significance":"If the reported performance differences persist across a broader set of instances, the paper would provide practitioners with a useful empirical comparison of a certified algebraic solver against a widely used numerical continuation solver. The strengths include the use of an independently developed external solver (HC) as a baseline, a reproducible repository, and the certified nature of the RUR solver, which avoids numerical non-convergence. However, the evidence base is narrow: one parameter set per model, no error bars or repeated runs, and the only examples favoring HC are artificially constructed. The significance is therefore conditional on the representativeness of the benchmark.","major_comments":[{"comment":"The abstract's claim 'and vice versa' — that homotopy methods can tackle examples out of reach for the RUR solver — is not supported by Table 1. In the rows where HC Julia is faster and more accurate (Crauste-2 and Crauste-3), the RUR solver still solves the systems; the paper itself labels these as 'More challenging for RUR,' not 'out of reach.' No row shows HC succeeding within the stated limits while RUR fails. The authors should either provide a genuine example where RUR does not terminate or fails, or revise the abstract and the concluding sentence of the introduction to state the actual finding: RUR extends the class of solvable instances on some problems, while on other problems HC is more efficient but RUR still succeeds.","section":"Section 2 (Main Result), Table 1 and bullet list"},{"comment":"The generalization of the benchmark conclusions rests on an untested assumption that 'the solvers behave similarly for different numerical values.' Since the polynomial systems are constructed from specific data, the number of solutions and the conditioning can vary substantially with the parameter values and observations. With a single parameter set per model, the 'out of reach' classification for Akt-2 and the relative performance ordering may be specific to the tested instances. The authors should test multiple parameter sets per model (or at least provide a sensitivity analysis) or explicitly restrict the claims to the tested data.","section":"Section 2 (Main Result), paragraph 'We tested one set of parameters per model'"},{"comment":"The only evidence for the HC-favorable direction is based on Crauste-2 and Crauste-3, which the authors acknowledge are artificially constructed by squaring parameters in Crauste-1 because they 'had difficulties to illustrate that with our small set of examples.' This is a significant weakness: the comparison does not demonstrate that the HC advantage occurs in natural parameter estimation problems. If the 'vice versa' claim is retained, the authors should find real ODE models where RUR is out of reach, or clearly state that the observed advantage is limited to a constructed class of systems.","section":"Section 2 (Main Result), bullet 'More challenging for RUR'"}],"minor_comments":[{"comment":"The table caption defines n/a, OOM, and timeout, but the body would be clearer if each n/a entry were accompanied by a reason (e.g., convergence failure, numerical blow-up) rather than leaving the reader to infer it from the text.","section":"Section 2, Table 1"},{"comment":"The 'relative error' column is not defined: it is unclear whether it is the maximum relative error over all estimated parameters, or the relative residual of the polynomial system. Please define the error metric.","section":"Section 2, Table 1"},{"comment":"The sentence 'We have discovered that such polynomial systems typically have very few solutions' is supported only by the toy example and the observed solution counts in Table 2. As stated, it is a heuristic claim; the paper should either provide more systematic evidence or temper the wording.","section":"Section 1, Introduction"},{"comment":"The comparison would be more reproducible if the exact version of HomotopyContinuation.jl and the chosen tolerance/precision settings were reported, since continuation solvers are sensitive to these parameters.","section":"Section 2, Experimental setup"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is very short and reads more like a benchmark report than a full methods paper. The RUR solver itself is described only by references to the authors' own packages, and the algorithmic contribution is not detailed. The central comparative claim is narrower than the abstract suggests. For a journal, the authors should either broaden the experimental evidence or reposition the paper as a focused empirical comparison."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper's real contribution is the integration of a certified RUR-based polynomial solver into the differential-algebra parameter estimation pipeline, plus a 12-model benchmark with code provided. That's a legitimate, useful extension over prior work, even though no new mathematics is introduced. The table is readable, the repo is public, and the authors are honest about having constructed Crauste-2 and Crauste-3 artificially because they couldn't find natural examples where HC clearly beats RUR. Credit where it's due: that's the right level of transparency. The soft spots are real but not fatal. The abstract's 'and vice versa' is not established: there is no row in Table 1 where HC solves a system that RUR cannot. In Crauste-2 and Crauste-3 HC is faster and more accurate, but RUR still solves them. The paper itself says 'more challenging for RUR,' not 'out of reach.' So the two-way claim should be one-way: RUR handles some systems HC can't finish, and on some systems HC is faster but RUR still succeeds. That's still interesting, just less dramatic. The one-parameter-set-per-model assumption is the biggest methodological weakness. The authors justify it by saying they expect similar behavior for different numerical values, but that's an assertion, not evidence. Solver performance on polynomial systems can depend heavily on coefficients, and since the systems are data-dependent, one run per model gives no sense of variance. Repeated runs with different parameter sets would be easy to add and would materially strengthen the benchmark. The timeout and memory limits (1 day, 100 GB) are arbitrary but reasonable; I don't think that's a flaw, just a boundary condition. The cases where HC returns no result (Akt-1, Goodwin, PK1, Treatment) are interesting but under-explored. The paper attributes them to structural non-identifiability causing numerical instability, but doesn't verify that mechanism. That's a minor concern; the observation itself stands. The comparison is against an independent external solver, HomotopyContinuation.jl, so the circularity burden is low. Self-citing their own RUR implementation is fine because the RUR solver is from their prior work and they're evaluating it, not claiming it as original here. Bottom line: this is a workmanlike benchmark paper with an overstated two-word phrase. With the 'vice versa' removed or replaced by a precise one-way claim, and with a few repeated runs, it's a solid contribution for people choosing polynomial solvers in this pipeline. It deserves a serious referee, not a desk reject, but the referee should push for the revision.","headline":"A genuinely useful benchmark with reproducible code, but the 'vice versa' claim in the abstract is not supported by the reported data; the paper deserves referee time but needs a tempered conclusion.","tokens_in":748,"tokens_out":1779,"would_cite":true,"duration_ms":32582,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A certified algebraic solver cracks parameter-estimation systems that homotopy continuation cannot finish.","keywords":["parameter estimation","ODE models","polynomial system solving","rational univariate representation","homotopy continuation","real root isolation","differential algebra"],"falsifier":"Take a second biologically plausible parameter set for the Akt-2 model and run both solvers under the same one-day limit: if the continuation solver finishes before the RUR solver, the claim that Akt-2 is out of reach for homotopy continuation is not universal.","tokens_in":3401,"feed_emoji":"🧮","tokens_out":7904,"duration_ms":68983,"temperature":0.7,"pith_summary":"The paper argues that the polynomial systems arising in differential-algebraic parameter estimation are typically much smaller in actual solution count than their Bezout bounds suggest, and that this makes certified solving practical. It compares two polynomial solvers as backends: a new solver based on rational univariate representation (RUR) with real root isolation, and the numerical homotopy continuation solver. On several biological ODE models the certified solver solves systems the continuation solver cannot finish, including a 69-unknown system with 80 solutions solved in about an hour; on two artificially symmetrized Crauste models the roles reverse. The takeaway is that the solver choice matters and neither approach dominates.","feed_headline":"Certified solver cracks parameter system homotopy can't finish","feed_subtitle":"A 69-unknown model with 80 true solutions took one hour; homotopy continuation ran more than a day without an answer.","key_machinery":"The load-bearing object is the rational univariate representation (RUR): a certified description of all solutions of a zero-dimensional polynomial system as rational functions of a single variable, produced from a lexicographic Groebner basis and combined with real root isolation. Because RUR certifies the solution set rather than tracking numerical paths, it is not disturbed by ill-conditioned or non-identifiable coordinate blow-ups; this is what lets it succeed where homotopy continuation times out or returns no solutions. The comparison baseline is homotopy continuation, which tracks solution paths from a generic start system to the target system; its speed depends on the number of paths, which scales with the Bezout bound, so systems with 80 actual solutions but astronomically large Bezout bounds remain slow or impractical. RUR instead pays for the actual solution count, which the paper observes is typically small for parameter-estimation systems.","core_discovery":"The central claim is that the new RUR solver can tackle examples that are out of reach for homotopy continuation methods, and vice versa. Concretely, the Akt-2 parameter estimation system has 69 unknowns, a Bezout bound of $2^{26}\\cdot 33^7$, but only 80 actual solutions; the RUR-based solver found them in about an hour, while the continuation solver did not finish within a day. The paper also reports cases where continuation returns no solutions on structurally non-identifiable models, while RUR finds them. The opposite direction is demonstrated by Crauste-2 and Crauste-3, artificially constructed by squaring parameters to introduce symmetries and increase the solution count from 32 to 128 and 512; there, homotopy continuation is faster and more accurate than RUR. The paper's conclusion is that a certified algebraic backend can be competitive and sometimes more robust on parameter estimation problems with models up to roughly ten states and ten parameters.","pith_inferences":["Editorial inference: if the small-actual-solution-count pattern holds broadly, a cheap structural pre-analysis could predict which backend will win, allowing a hybrid pipeline to be scheduled automatically.","Editorial inference: the RUR solver's slower, less accurate runs on Crauste-2 and Crauste-3 suggest that a hybrid strategy—certify with RUR, then refine with homotopy—could combine reliability with speed.","Editorial inference: because only one parameter set per model was tested, backend rankings may shift on other data; a natural extension is a parameter sweep measuring how often each solver wins."],"forward_implications":["For parameter-estimation systems with few actual solutions, certified RUR solving can replace long or unsuccessful numerical continuation runs with a one-hour computation and a guaranteed answer.","Structurally non-identifiable states or parameters, which often destabilize numerical path tracking, can be handled by the certified solver where continuation returns no solutions.","On systems with many solutions created by symmetries, homotopy continuation remains the faster and more accurate backend, so the practical solver choice depends on structural features of the model.","The observed gap between Bezout bounds and actual solution counts makes algebraic certification feasible for a wider range of biological models than previously assumed."],"supporting_citations":[{"why":"Constructs the differential-algebra plus rational-interpolation pipeline that turns parameter estimation into polynomial system solving.","marker":"(Bassik et al., 2023)"},{"why":"Provides the homotopy continuation solver used as the baseline comparison throughout the paper.","marker":"(Breiding and Timme, 2018)"},{"why":"Supplies the benchmark ODE models and the identifiability classification used to explain cases where continuation fails.","marker":"(Barreiro and Villaverde, 2023)"},{"why":"Foundational description of the rational univariate representation solver on lexicographic Groebner bases that the new backend implements.","marker":"(Demin et al., 2024)"},{"why":"Source of the Crauste-1 CD8 T-cell model from which the symmetric Crauste-2 and Crauste-3 cases are constructed.","marker":"(Crauste et al., 2017)"}],"fun_headline_variants":["RUR solver beats homotopy on 69-unknown parameter system","Certified solver finds 80 solutions in an hour that homotopy missed","RUR beats homotopy: solves 69 unknown model in 1 hour vs 1 day","Homotopy stalls, certified solver wins: 80 solutions in an hour","Certified polynomial solver solves in 1 hour what homotopy can't in a day"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The benchmark treats one tested parameter set per model as representative of solver behavior, so the reported winners and losers may not generalize to other parameter values or data sets.","fun_headline_variants_meta":{"raw":{"variants":["RUR solver beats homotopy on 69-unknown parameter system","Certified solver finds 80 solutions in an hour that homotopy missed","RUR beats homotopy: solves 69 unknown model in 1 hour vs 1 day","Homotopy stalls, certified solver wins: 80 solutions in an hour","Certified polynomial solver solves in 1 hour what homotopy can't in a day"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001132,"raw_usage":{"total_tokens":4666,"prompt_tokens":867,"completion_tokens":3799,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":483,"completion_tokens_details":{"reasoning_tokens":3691}},"tokens_in":483,"tokens_out":3799,"duration_ms":22044,"temperature":1.0,"reasoning_tokens":3691,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:44:31.430146+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a second biologically plausible parameter set for the Akt-2 model and run both solvers under the same one-day limit: if the continuation solver finishes before the RUR solver, the claim that Akt-2 is out of reach for homotopy continuation is not universal.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Constructs the differential-algebra plus rational-interpolation pipeline that turns parameter estimation into polynomial system solving."},{"cited_title":"and Timme, S","cited_arxiv_id":null,"evidence_quote":"Provides the homotopy continuation solver used as the baseline comparison throughout the paper."},{"cited_title":"and Villaverde, A.F","cited_arxiv_id":null,"evidence_quote":"Supplies the benchmark ODE models and the identifiability classification used to explain cases where continuation fails."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source of the Crauste-1 CD8 T-cell model from which the symmetric Crauste-2 and Crauste-3 cases are constructed."}],"review_version":1}