Reaction precedents extend retrosynthetic prediction beyond training coverage

Jiaxing Sun; Xuli Shen; Jiang Wu; Ruiyi Ding; Ruikang Zhang; Xiao An; Xibao Li; Yubin Wang

cs.AIOct 29, 2026

Reaction precedents extend retrosynthetic prediction beyond training coverage

Authors: Jiaxing Sun, Xuli Shen, Jiang Wu, Ruiyi Ding, Ruikang Zhang, Xiao An, Xibao Li, Yubin Wang,

Organizations: Shanghai Artificial Intelligence Laboratory · Fudan University · Peking University · Wuhan University ·

Abstract

Reaction precedents guide retrosynthesis, but how far they can compensate for gaps in task-specific training coverage remains unclear. Using 2.71 million reactions, we restricted training and retrieval coverage by transformation or time, then expanded only the retrieval corpus with model parameters fixed. For transformations excluded from training, mean Top-1 accuracy increased from 13.0% to 58.0% on a historical test set and from 4.9% to 23.1% on a temporally held-out test set. For temporal coverage gaps, retrieval expansion enabled a model trained on records before 2017 to match or exceed a non-retrieval model trained on records before 2025 on both test sets. Without task-specific fine-tuning, precedents improved all 14 representative general-purpose models on both test sets and USPTO-50K; the strongest with precedents surpassed all eight representative specialized models. These findings show that inference-time reaction precedents can extend retrosynthetic prediction beyond task-specific training coverage without parameter updates.

Figures & tables

Figure 2

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Partition	Dataset identifier	Recorded years	Reactions	Role
Train-Full	DianShi-PatentRxnStep-Train-Full	1976–2024	2,712,259	Full training partition; also supplies the full retrieval corpus
Val-2001-2024	DianShi-PatentRxnStep-Val-2001-2024	2001–2024	3,798	Validation on historical records
Test-2001-2024	DianShi-PatentRxnStep-Test-2001-2024	2001–2024	3,798	Evaluation on historical records
Val-2025	DianShi-PatentRxnStep-Val-2025	2025	1,886	Temporally held-out validation
Test-2025	DianShi-PatentRxnStep-Test-2025	2025	1,886	Temporally held-out evaluation

Table 10

Model	Training corpus	Inference retrieval corpus	Comparison role
RetroStepLM-Direct	TemplateRestricted-S{s}	None	Separately trained template-restricted reference
RetroStepLM-RAG	TemplateRestricted-S{s}	TemplateRestricted-S{s}	Restricted-retrieval condition
Same RetroStepLM-RAG checkpoint	TemplateRestricted-S{s}	Full	Expanded-retrieval condition
RetroStepLM-Direct	Pre2017	None	Separately trained temporally restricted reference
RetroStepLM-RAG	Pre2017	Pre2017	Restricted-retrieval condition
Same RetroStepLM-RAG checkpoint	Pre2017	Full	Expanded-retrieval condition

Table 11

Processing component	Rule	Scope
Molecular standardization	Standardize reactant and product SMILES using RDKit	Reaction records entering dataset construction
Reaction deduplication	Compare standardized reactant-set–product-set signatures after removing mapping numbers and sorting components; retain the earliest recorded year for cross-year duplicates	Duplicate reaction records
Structural quality filtering	Exclude multiple products, repeated molecules, reactant–product overlap, more than four reactant components and inorganic products	Deduplicated records
Mapping and extraction filtering	Exclude inconsistent atom mappings and unsuccessful reaction-centre or template extraction	Records undergoing quality control
Template-frequency filtering	Remove local reaction templates represented by fewer than six reactions	Quality-controlled corpus, before temporal partitioning
External benchmark exclusions	Compare standardized target products against the specified external dataset partitions	Training and evaluation candidates, as detailed in section 9.2

Table 12

Item	Setting
Base model	Qwen3-4B-Base
Parameter updates	Full-parameter supervised fine-tuning
Supervision	Synthon–leaving-group intermediate labels
Hardware	32 H200 GPUs across four nodes
Precision	bfloat16
Training implementation	FlashAttention; DeepSpeed ZeRO-3; gradient checkpointing

Table 13

Record set	Reactions retained	Percentage of Train-Full (%)	Reactions removed
Train-Full	2,712,259	100.00	0
TemplateRestricted-S42	1,804,926	66.55	907,333
TemplateRestricted-S43	1,673,416	61.70	1,038,843
TemplateRestricted-S44	1,564,454	57.68	1,147,805

Table 14

Partition	2001–2024 / Retained	2001–2024 / Withheld	2025 / Retained	2025 / Withheld
S42	1,977	1,821	970	916
S43	1,842	1,956	937	949
S44	1,797	2,001	878	1,008

Table 15

Partition	Configuration	Top-1	Top-3	Top-5	Top-10
S42	RetroStepLM-Direct	67.32	79.46	83.11	86.75
S42	RetroStepLM-RAG , restricted retrieval	72.38	84.93	88.42	91.20
S42	RetroStepLM-RAG , full retrieval	71.78	84.88	87.96	91.00
S43	RetroStepLM-Direct	66.56	78.34	81.16	84.91
S43	RetroStepLM-RAG , restricted retrieval	69.00	81.65	84.96	87.35
S43	RetroStepLM-RAG , full retrieval	66.94	81.32	85.12	88.00

Table 16

Partition	Configuration	Top-1	Top-3	Top-5	Top-10
S42	RetroStepLM-Direct	4.39	11.26	15.54	21.25
S42	RetroStepLM-RAG , restricted retrieval	11.81	25.65	33.66	46.07
S42	RetroStepLM-RAG , full retrieval	56.07	71.00	75.23	79.96
S43	RetroStepLM-Direct	4.24	10.12	14.57	21.47
S43	RetroStepLM-RAG , restricted retrieval	14.37	27.66	34.92	45.76
S43	RetroStepLM-RAG , full retrieval	58.08	72.03	76.23	81.19

Table 17

Partition	Configuration	Top-1	Top-3	Top-5	Top-10
S42	RetroStepLM-Direct	41.24	57.63	62.27	67.11
S42	RetroStepLM-RAG , restricted retrieval	40.62	58.35	65.46	70.62
S42	RetroStepLM-RAG , full retrieval	37.94	57.53	63.51	69.69
S43	RetroStepLM-Direct	38.53	54.43	61.58	69.05
S43	RetroStepLM-RAG , restricted retrieval	37.46	51.97	57.10	62.86
S43	RetroStepLM-RAG , full retrieval	33.94	48.13	53.90	60.51

Table 18

Partition	Configuration	Top-1	Top-3	Top-5	Top-10
S42	RetroStepLM-Direct	3.71	8.52	12.55	18.78
S42	RetroStepLM-RAG , restricted retrieval	3.71	11.68	18.56	29.15
S42	RetroStepLM-RAG , full retrieval	20.96	32.53	38.32	46.07
S43	RetroStepLM-Direct	2.42	7.06	12.01	18.86
S43	RetroStepLM-RAG , restricted retrieval	5.27	14.96	20.34	29.50
S43	RetroStepLM-RAG , full retrieval	25.29	38.15	41.41	47.52

Table 19

Reactions per template in full corpus	Partition	Distinct templates	Evaluation reactions	RetroStepLM-Direct	RetroStepLM-RAG , restricted retrieval	RetroStepLM-RAG , full retrieval
1–100	S42	290	305	6.56	8.85	48.52
1–100	S43	290	309	5.18	11.97	47.90
1–100	S44	290	308	5.84	14.29	48.70
101–1,000	S42	269	446	4.71	11.88	53.36
101–1,000	S43	269	452	5.53	11.50	54.42
101–1,000	S44	269	429	5.36	15.15	54.31

Table 20

Reactions per template in full corpus	Partition	Distinct templates	Evaluation reactions	RetroStepLM-Direct	RetroStepLM-RAG , restricted retrieval	RetroStepLM-RAG , full retrieval
1–100	S42	188	212	4.25	4.72	11.32
1–100	S43	188	204	2.94	3.43	11.27
1–100	S44	188	214	5.14	3.27	11.68
101–1,000	S42	147	222	2.70	1.35	20.27
101–1,000	S43	147	211	2.37	3.32	27.49
101–1,000	S44	147	235	2.55	5.96	17.45

Table 21

Model	Training corpus	Inference retrieval corpus	Top-1	Top-3	Top-5	Top-10
RetroStepLM-Direct	Pre2017	None	64.2	79.8	84.0	87.7
RetroStepLM-Direct	Full	None	64.3	80.3	85.0	89.2
RetroStepLM-RAG	Pre2017	Pre2017	73.5	88.4	92.1	94.8
RetroStepLM-RAG	Pre2017	Full	73.3	88.5	92.0	94.7
RetroStepLM-RAG	Full	Full	72.8	88.6	92.2	94.9

Table 22

Model	Training corpus	Inference retrieval corpus	Top-1	Top-3	Top-5	Top-10
RetroStepLM-Direct	Pre2017	None	50.8	67.3	71.6	76.3
RetroStepLM-Direct	Full	None	62.0	76.7	80.9	84.3
RetroStepLM-RAG	Pre2017	Pre2017	58.6	74.3	79.9	84.2
RetroStepLM-RAG	Pre2017	Full	70.7	84.4	88.6	91.5
RetroStepLM-RAG	Full	Full	72.4	86.4	89.8	92.7

Table 23

Model	Training corpus	Inference retrieval corpus	Top-1	Top-3	Top-5	Top-10
RetroStepLM-Direct	Pre2017	None	25.8	41.0	46.5	52.5
RetroStepLM-Direct	Full	None	36.4	53.6	59.4	64.7
RetroStepLM-RAG	Pre2017	Pre2017	28.7	45.4	52.7	61.8
RetroStepLM-RAG	Pre2017	Full	36.5	53.8	61.4	69.0
RetroStepLM-RAG	Full	Full	39.4	59.1	67.0	74.9

Table 24

Condition	Selection rule
Standard retrieval	Select precedents by target-product structural similarity, without reference reaction-class or template constraints
Reference-class restriction	Select precedents sharing the recorded reference reaction class
Reference-template restriction	Select precedents sharing the local reaction template of the reference reaction
Matching-template exclusion	Exclude precedents bearing the local reaction template of the reference reaction

Table 25

Model	No precedents	Standard retrieval	Reference-class restriction	Reference-template restriction	Matching-template exclusion
gemini-3.8-flash	42.94	70.80	74.16	90.75	45.24
deepseek-v4.1-flash	32.63	62.05	71.44	92.93	20.77
Qwen3-32B	0.00	20.61	45.86	60.52	2.70
Qwen3.5-27B	1.08	36.85	54.16	70.36	5.45
RetroStepLM-RAG	—	72.80	72.16	79.33	56.22

Table 26

Model	No precedents	Standard retrieval	Reference-class restriction	Reference-template restriction	Matching-template exclusion
gemini-3.8-flash	27.49	69.33	73.35	88.34	37.15
deepseek-v4.1-flash	19.19	60.87	71.20	90.23	17.67
Qwen3-32B	0.05	22.67	46.74	58.95	2.79
Qwen3.5-27B	0.26	36.12	52.40	67.11	4.71
RetroStepLM-RAG	—	72.40	71.77	78.23	47.10

Table 27

Model	No precedents	Standard retrieval	Reference-class restriction	Reference-template restriction	Matching-template exclusion
gemini-3.8-flash	22.85	35.74	46.45	73.49	19.35
deepseek-v4.1-flash	15.85	30.70	44.80	77.15	8.43
Qwen3-32B	0.00	11.56	20.62	34.52	1.27
Qwen3.5-27B	0.27	13.52	25.50	44.33	1.64
RetroStepLM-RAG	—	39.40	43.11	50.11	30.01

Table 28

Evaluation group	Model identifiers
API-accessed models	glm-5.2 ; qwen3.8-MAX ; kimi-k3 ; deepseek-v4-pro ; deepseek-v4.1-flash ; gpt-5.6-sol ; gemini-3.8-flash
Open-weight Qwen models	Qwen3-4B ; Qwen3-8B ; Qwen3-32B ; Qwen3.5-2B ; Qwen3.5-4B ; Qwen3.5-9B ; Qwen3.5-27B

Table 29

Item	Setting
Temperature	0
Top-p	1.0
Maximum output tokens	65,535
Thinking	Provider-default settings
Incomplete-response handling	Rerun the API request until a complete response was obtained
Scored output	One final precursor set per target

Table 30

Condition	Retrieval record set	Reactions	Precedents per query
NoPrecedents	None	0	0
RXNGraphormer corpus	RXNGraphormer-RetrievalCorpus , constructed from the RXNGraphormer training corpus	6,591,946	5
USPTO corpus	USPTO-RetrievalCorpus , assembled by merging the training sets of USPTO-50K , USPTO-MIT and USPTO-Full	1,217,980	5
DianShi corpus	DianShi-PatentRxnStep-Train-Full	2,712,259	5

Table 31

Model	NoPrecedents	RXNGraphormer corpus	USPTO corpus	DianShi corpus
glm-5.2	20.17	38.89	46.24	53.45
qwen3.8-MAX	21.69	37.55	44.70	53.43
kimi-k3	34.39	49.83	54.70	59.94
deepseek-v4-pro	29.38	35.85	49.97	60.14
deepseek-v4.1-flash	32.63	41.92	53.33	62.05
gpt-5.6-sol	41.42	42.34	54.54	63.89

Table 32

Model	NoPrecedents	RXNGraphormer corpus	USPTO corpus	DianShi corpus
glm-5.2	9.98	20.64	22.22	52.79
qwen3.8-MAX	10.80	20.06	22.12	52.87
kimi-k3	17.98	27.96	28.36	58.58
deepseek-v4-pro	16.03	19.48	25.28	59.69
deepseek-v4.1-flash	19.19	22.56	27.80	60.87
gpt-5.6-sol	25.09	24.54	29.91	63.95

Table 33

Model	NoPrecedents	RXNGraphormer corpus	USPTO corpus	DianShi corpus
glm-5.2	8.17	9.17	11.03	21.84
qwen3.8-MAX	8.75	9.76	10.82	21.58
kimi-k3	15.48	15.32	17.66	27.09
deepseek-v4-pro	13.20	10.13	15.06	29.90
deepseek-v4.1-flash	15.85	11.98	15.32	30.70
gpt-5.6-sol	19.19	11.93	15.91	30.12

Table 34

Component	Protocol
Model parameters	Released weights
Retraining on DianShi data	None
Input formatting	According to each method’s requirements
Molecular test-time augmentation on Test-2001-2024 and Test-2025	None
Scoring	Common standardized complete precursor-set exact match
Reported ranks	Top-1, Top-3, Top-5 and Top-10

Table 35

Model	Top-1	Top-3	Top-5	Top-10
LocalRetro	53.4	77.5	85.9	92.4
Graph2Edits	56.3	79.2	86.2	91.0
R-SMILES	55.1	77.3	83.4	89.4
EditRetro	60.8	80.6	86.0	90.3
NAG2G	55.1	76.9	83.4	89.9
ReactSeq	58.9	80.5	86.4	91.4

Table 36

Model	Top-1	Top-3	Top-5	Top-10
LocalRetro	22.7	35.4	40.8	45.7
Graph2Edits	22.1	33.9	38.8	42.7
R-SMILES	23.8	36.0	41.0	45.4
EditRetro	19.2	29.9	31.8	32.2
NAG2G	23.2	36.0	40.7	45.9
ReactSeq	22.8	35.4	39.7	43.9

Table 37

Model	Top-1	Top-3	Top-5	Top-10
LocalRetro	18.3	29.1	33.7	38.0
Graph2Edits	16.4	26.9	31.3	36.4
R-SMILES	17.0	28.3	32.8	37.6
EditRetro	14.8	22.5	23.8	24.0
NAG2G	16.4	27.8	33.1	38.0
ReactSeq	17.4	27.6	32.3	37.0

Table 38

Benchmark	Source	Targets	Average steps
Pistachio Hard	Public benchmark derived from Pistachio reaction data	100	7.2 38
USPTO-190	Public benchmark introduced with Retro*	190	6.0 38
PatMol-150	This study; 2025 public patents	150	8.17
The minimum reference-route length for PatMol-150 is 7 steps.

Table 39

Maximum reaction depth	Routes	Percentage (%)
7	63	42.0
8	46	30.7
9	17	11.3
10	9	6.0
11	9	6.0
12	3	2.0

Table 40

Component	Setting
Template preselection	Product PatternFP; Tversky similarity 1 ( $\alpha=0$ , $\beta=1$ )
Structural matching	One-to-one SMARTS matching on product and reactant sides
Initial template application	Up to 256 matching templates rewritten with rdchiral
Template-search fallback	1,000 highest reactant-side fingerprint matches among product-preselected templates
Fallback ranking	AOT*-style reward on reactant sets; best retained; $\gamma=0.1$
Unsolved penalty	Applied only on the template-search fallback path

Table 41

Single-step model	Template-library source	Entries
RetroStepLM-Direct and RetroStepLM-RAG	Train-Full	412,027
RSGPT and RXNGraphormer	DESP USPTO library	270,794
RetroAgent default model	DESP USPTO library	270,794
Retro-R1 default model	Retro* USPTO library	381,302

Table 42

Framework	Budget unit
RetroAgent	Newly expanded distinct molecules
Retro-R1-modified	Single-step tool calls
DianShi-RetroPlanSystem-v0	Expanded molecule nodes; each expansion invoked the single-step model once

Table 43

Benchmark	Single-step model	$N=50$	$N=100$	$N=200$	$N=300$	$N=400$	$N=500$
Pistachio Hard	Default	56.00	56.00	56.00	56.00	56.00	56.00
Pistachio Hard	RXNGraphormer	47.00	47.00	47.00	47.00	47.00	47.00
Pistachio Hard	RSGPT	76.00	77.00	79.00	79.00	79.00	79.00
Pistachio Hard	RetroStepLM-Direct	91.00	92.00	92.00	92.00	92.00	92.00
Pistachio Hard	RetroStepLM-RAG	86.00	90.00	90.00	90.00	90.00	90.00
USPTO-190	Default	37.89	38.95	38.95	38.95	38.95	38.95

Table 44

Benchmark	Single-step model	$N=50$	$N=100$	$N=200$	$N=300$	$N=400$	$N=500$
Pistachio Hard	Default	69.00	69.00	69.00	69.00	69.00	69.00
Pistachio Hard	RXNGraphormer	35.00	36.00	36.00	36.00	36.00	36.00
Pistachio Hard	RSGPT	78.00	78.00	79.00	79.00	79.00	79.00
Pistachio Hard	RetroStepLM-Direct	84.00	84.00	84.00	84.00	84.00	84.00
Pistachio Hard	RetroStepLM-RAG	91.00	92.00	92.00	92.00	92.00	92.00
USPTO-190	Default	43.16	43.16	43.16	43.16	43.16	43.16

Table 45

Benchmark	Single-step model	$N=50$	$N=100$	$N=200$	$N=300$	$N=400$	$N=500$
Pistachio Hard	RXNGraphormer	39.00	40.00	44.00	44.00	44.00	44.00
Pistachio Hard	RSGPT	76.00	78.00	86.00	88.00	88.00	88.00
Pistachio Hard	RetroStepLM-Direct	88.00	88.00	90.00	93.00	94.00	94.00
Pistachio Hard	RetroStepLM-RAG	89.00	91.00	94.00	95.00	96.00	96.00
USPTO-190	RXNGraphormer	27.37	32.11	34.74	35.79	35.79	36.32
USPTO-190	RSGPT	61.58	65.79	81.05	82.63	83.68	83.68

Table 46

Training corpus	Intermediate labels	Retrieval	Top-1	Top-3	Top-5	Top-10
None	No	No	0.0	0.0	0.0	0.0
Train-Full	No	No	61.49	78.37	82.98	86.74
Train-Full	Yes	No	64.3	80.3	85.0	89.2
Train-Full	No	Yes	71.70	86.84	90.71	94.19
Train-Full	Yes	Yes	72.8	88.6	92.2	94.9

Table 47

Training corpus	Intermediate labels	Retrieval	Top-1	Top-3	Top-5	Top-10
None	No	No	0.0	0.0	0.0	0.0
Train-Full	No	No	58.29	73.85	78.07	82.10
Train-Full	Yes	No	62.0	76.7	80.9	84.3
Train-Full	No	Yes	70.06	84.81	88.47	91.73
Train-Full	Yes	Yes	72.4	86.4	89.8	92.7

Table 48

Training corpus	Intermediate labels	Retrieval	Top-1	Top-3	Top-5	Top-10
None	No	No	0.0	0.0	0.0	0.0
Train-Full	No	No	35.63	51.48	57.26	63.26
Train-Full	Yes	No	36.4	53.6	59.4	64.7
Train-Full	No	Yes	34.36	54.19	61.40	69.19
Train-Full	Yes	Yes	39.4	59.1	67.0	74.9

Table 49

Training-time selection	Top-1	Top-3	Top-5	Top-10
Deterministic Tanimoto Top-5	71.66	87.20	91.33	94.21
Random five from Tanimoto Top-10	72.8	88.6	92.2	94.9

Table 50

Training-time selection	Top-1	Top-3	Top-5	Top-10
Deterministic Tanimoto Top-5	70.19	84.99	88.84	92.18
Random five from Tanimoto Top-10	72.4	86.4	89.8	92.7

Table 51

Training-time selection	Top-1	Top-3	Top-5	Top-10
Deterministic Tanimoto Top-5	36.80	54.88	62.25	70.15
Random five from Tanimoto Top-10	39.4	59.1	67.0	74.9

Table 52

Training history	Top-1	Top-3	Top-5	Top-10
Train-Full	64.3	80.3	85.0	89.2
Train-Full $\rightarrow$ USPTO-50K-train	67.9	86.0	90.9	95.0

Table 53

Training history	Top-1	Top-3	Top-5	Top-10
Train-Full	62.0	76.7	80.9	84.3
Train-Full $\rightarrow$ USPTO-50K-train	57.4	76.2	81.7	86.5

Table 54

Training history	Top-1	Top-3	Top-5	Top-10
Train-Full	36.4	53.6	59.4	64.7
Train-Full $\rightarrow$ USPTO-50K-train	34.6	52.9	60.9	69.5

Table 55

Inference retrieval corpus	Top-1	Top-3	Top-5	Top-10
USPTO-50K-train	53.88	73.66	80.27	85.98
Pre2017	64.81	81.85	86.90	91.13
Full	64.57	81.33	87.00	91.31

Table 56

Inference retrieval corpus	Top-1	Top-3	Top-5	Top-10
USPTO-50K-train	20.75	31.49	35.49	40.44
Pre2017	37.52	49.87	53.98	58.48
Full	44.92	57.82	61.90	66.14

Table 57

Inference retrieval corpus	Top-1	Top-3	Top-5	Top-10
USPTO-50K-train	14.48	23.17	27.68	31.97
Pre2017	16.44	26.56	31.28	36.00
Full	21.16	32.24	37.01	42.05

Table 58

Retrosynthesis

See all 25 papers

Single-step Retrosynthesis via Reinforcement Learning from Forward Feedback
Oct 10, 2026Zixun Lan, Zhenfu Liu, Yi Li +6Retrosynthesis Reinforcement Learning
Xi’an Jiaotong-Liverpool University, Suzhou 215000, China · Sichuan University of Science & Engineering, Zigong 643000, China · Fudan University, Shanghai 200433, China +1
Coverage Before Control: Route-Instruction Grounding and Steering for Controllable Retrosynthesis
Sep 30, 2026Xuemin Chen, Xiaozhuang Song, Xinjian Zhao +2Instruction Following Instruction Tuning
The Chinese University of Hong Kong, Shenzhen
Training Chemical Plausibility-Aware Large Language Models for Single-Step Retrosynthesis
Aug 19, 2026Bogdan Zagribelnyy, Ivan Ilin, Nikita Bondarev +5Chemical Reaction Prediction Retrosynthesis
Insilico Medicine AI Limited, Abu Dhabi, UAE · Insilico Medicine Canada Inc., Montreal, Quebec, Canada · Insilico Medicine Hong Kong Ltd., Hong Kong SAR, China

Oct 10, 2026J/K move · Enter open · S save

Single-step Retrosynthesis via Reinforcement Learning from Forward Feedback

Zixun Lan, Zhenfu Liu, Yi Li, Xinheng Wang+5

Xi’an Jiaotong-Liverpool University, Suzhou 215000, China · Sichuan University of Science & Engineering, Zigong 643000, China · Fudan University, Shanghai 200433, China+1

Retrosynthesis Reinforcement Learning

Open paper AskPDF arXiv

Single-step retrosynthesis is a one-to-many inverse problem, yet most models are trained and evaluated against a single recorded reactant set. Reference-based training gives no explicit positive signal for unrecorded alternatives, and exact-match evaluation rejects non-matching predictions regardless of chemical plausibility. We propose Retrosynthesis via Reinforcement Learning from Forward Feedback (RetroRLFF), which optimizes reactant generation using both reference agreement and forward consistency with the target product. Starting from a supervised decoder-only Transformer, RetroRLFF generates complete reactant candidates and uses a frozen forward reaction model to predict products. A sequence-level reward combines reference agreement and forward consistency while penalizing invalid outputs. Direct Advantage-Based Policy Optimization updates the policy without human preference annotations or a preference-trained reward model. On USPTO-50K, RetroRLFF outperforms all 14 baselines in candidate-level round-trip accuracy at k = 1, 3, 5, and 10. With approximately 50 million parameters, it achieves 90.1% top-1 round-trip accuracy, exceeding R-SMILES, the strongest competing baseline on this metric, by 8.5 percentage points. Compared with its supervised policy, round-trip accuracy increases from 83.1% to 90.1%, while exact-match accuracy is maintained at 57.9% versus 57.5%. In an anonymized expert assessment of 131 randomly selected test cases, organic synthesis experts prefer RetroRLFF over EditRetro in 67.9% of comparisons. These results support forward feedback as a training signal for improving forward consistency while retaining top-1 reference recovery. Expert assessment provides complementary evidence of synthetic feasibility on the selected cases, although neither forward consistency nor expert preference establishes experimental feasibility.

Abstract

Retrosynthesis

Explore similar work

Single-step Retrosynthesis via Reinforcement Learning from Forward Feedback

Coverage Before Control: Route-Instruction Grounding and Steering for Controllable Retrosynthesis

Training Chemical Plausibility-Aware Large Language Models for Single-Step Retrosynthesis