---
title: Condition-Degradation Guidance (CDG)
url: https://www.emergentmind.com/topics/condition-degradation-guidance-cdg
type: topic
---

# Condition-Degradation Guidance (CDG)

Searching arXiv for the cited CDG-related papers to ground the article in the latest metadata.
Condition-Degradation Guidance (CDG) is a field-dependent research term that denotes guidance by degraded or condition-derived signals rather than a single canonical algorithm. In current arXiv usage, it appears in at least four distinct instantiations: a robust operations-and-maintenance framework in which streaming condition signals parameterize degradation models and uncertainty sets for multi-asset production and maintenance optimization [2311.10966]; a text-to-image diffusion guidance method that replaces the null prompt in Classifier-Free Guidance with a semantically degraded condition $\boldsymbol{c}_{\text{deg}}$ [2603.10780]; a power-system reliability framework for generating close-to-real in-group asset condition data from degradation, correlation, and categorical models [2011.12352]; and a POMDP-based operating-condition optimization methodology in which real-time multi-sensor degradation signals guide capacity and preventive-maintenance decisions under partial observability [2512.06682]. This multiplicity of meanings is central to the term: across domains, CDG consistently denotes the use of degraded, condition-conditioned, or condition-derived information to steer downstream decisions, but the modeled objects, mathematical structures, and objectives differ substantially.

## 1. Terminological scope and research domains

The term has explicit and implicit usages. In the diffusion-model literature, CDG is explicitly defined as a plug-and-play replacement for the null-prompt–based negative in CFG, with the degraded condition constructed directly from the positive prompt [2603.10780]. In power-system reliability assessment, CDG is explicitly defined as a guidance framework for generating synthetic in-group asset condition data when inspection records are unavailable or incomplete [2011.12352]. By contrast, in multi-asset O\&M optimization and in POMDP-based operating-condition control, CDG is not the paper’s explicit term; rather, it is an inferred conceptual alignment describing the closure of the loop between predictive monitoring and prescriptive decision-making [2311.10966] and the use of real-time multi-sensor degradation signals to guide operating-condition control and preventive maintenance [2512.06682].

| Domain | Meaning of CDG | Core mechanism |
|---|---|---|
| Multi-asset O\&M | Condition signals guiding robust production and maintenance | Robust MILP with degradation interactions |
| Diffusion models | Semantically degraded negative condition | Replace $\emptyset$ with $\boldsymbol{c}_{\text{deg}}$ in guidance |
| Power-system reliability | Generation of asset condition data from degradation guidance | Degradation, correlation, categorical, and probabilistic models |
| POMDP control | Real-time degradation signals guiding control and PM | Constrained IOHMM + belief-state optimization |

A common misconception is that CDG names a single mature methodology. The literature instead shows a polysemous term whose shared intuition is guidance by structured degradation information. This suggests a family resemblance rather than a unified formalism.

## 2. CDG in robust multi-asset operations and maintenance

In robust condition-based O\&M, CDG denotes the joint use of condition signals and explicit degradation models to co-optimize throughput, cost, and reliability in multi-asset systems with degradation interactions [2311.10966]. The framework defines assets $i \in A$, time periods $t \in T$, and, when multiple maintenances are allowed, maintenance cycles $k \in K$. Core decision variables include production $p_{i,t} \ge 0$, maintenance unavailability $u^m_{i,t}$, failure unavailability $u^f_{i,t}$, total unavailability $u_{i,t}=u^m_{i,t}+u^f_{i,t}$, preventive-maintenance starts $m^p_{i,t}$, corrective-maintenance starts $m^c_{i,t}$, degradation signal amplitude $l_{i,t}$, interaction surrogate $\omega_{i,t}$, operations-induced degradation coefficient $\zeta_{i,t}$, and multi-asset degradation interaction coefficient $\gamma_{j,i,t}$.

The continuous-time degradation model is decision-dependent and interaction-coupled:
$$
l_i(t)=l_i(0)+\int_0^t\left[D_i+\zeta_i(\rho_i(s))+\sum_{j\in A_i}\gamma_{j,i}(l_j(s))\right]ds+B_i(t).
$$
Its optimization embedding is a discrete-time linear constraint:
$$
l_{i,t} \ge l_{i,0} + \sum_{\tau=1}^{t} \left[d_{i,\tau}+\zeta_{i,\tau}p_{i,\tau}+\sum_{j\in A_i}\gamma_{j,i,\tau}\omega_{j,\tau}\right] - M\sum_{\tau=1}^{t}u_{i,\tau},
$$
with hard threshold enforcement $l_{i,t}\le \Lambda_i$. The $\zeta_{i,t}p_{i,t}$ term models operations-induced degradation (OID), while $\sum_{j\in A_i}\gamma_{j,i,t}\omega_{j,t}$ models multi-asset degradation interactions (MDI). The interaction surrogate $\omega_{i,t}$ equals $l_{i,t}$ except when an asset is failed, in which case it is pushed to $\Lambda_i$, or when it is under planned preventive maintenance, in which case there is no interaction.

A distinctive feature is the use of sensor-adaptive nested budgeted uncertainty sets. Posterior means $\bar d_{i,\tau}$, $\bar\zeta_{i,\tau}$, and $\bar\gamma_{j,i,\tau}$ and half-widths $\hat d_{i,\tau}$, $\hat\zeta_{i,\tau}$, and $\hat\gamma_{j,i,\tau}$ are obtained from observed degradation signals and embedded in condition-based uncertainty sets $U_{i,t}$, with $\Delta_{i,t}$ controlling conservatism. The objective minimizes preventive and corrective maintenance cost, failure-related cost, production cost, and unmet-demand penalty under throughput, capacity, and crew constraints. Robust degradation accumulation over maintenance cycles is dualized, producing a mixed-integer linear program with robust linear constraints.

The reported experiments compare four policies—Base, OID-only, MDI-only, and Comprehensive (OID+MDI)—under both emulated degradation and vibration-based readings from a rotating machinery system. In simulation averages, the deterministic comprehensive model improves O\&M cost by $50.8\%$, $23.5\%$, and $41.1\%$ versus Base, OID-only, and MDI-only, respectively; yields the lowest penalty cost for unsatisfied demand, approximately $5{,}097$ versus approximately $29{,}882$, $11{,}270$, and $28{,}048$; and reduces failures per horizon to approximately $0.86$ versus approximately $7.83$, $3.22$, and $4.13$. As the uncertainty budget $\Delta$ grows, the robust objective increases and aligns with simulation averages, with the best trade-off around $\Delta \approx 4$. An acceleration method—ultra-conservative warm start, cut generation, and robust MILP solution—improves solved-instance count and reduces average solution time and optimality gaps.

This instantiation of CDG is prescriptive rather than generative. Condition signals do not merely forecast failure; they parameterize the feasible set and the worst-case degradation scenarios that determine production rates, maintenance timing and type, and resource allocation.

## 3. CDG in diffusion-model guidance

In diffusion models, CDG is a guidance paradigm that replaces the semantically vacuous null prompt $\emptyset$ in Classifier-Free Guidance with a strategically degraded condition $\boldsymbol{c}_{\text{deg}}$ [2603.10780]. Standard CFG uses
$$
D_\theta^{\mathrm{CFG}}(\boldsymbol{x}_\sigma;\sigma,\boldsymbol{c})=
D_\theta(\boldsymbol{x}_\sigma;\sigma,\boldsymbol{c})
+(w-1)\big(D_\theta(\boldsymbol{x}_\sigma;\sigma,\boldsymbol{c})-D_\theta(\boldsymbol{x}_\sigma;\sigma,\emptyset)\big),
$$
whereas CDG replaces $\emptyset$ by $\boldsymbol{c}_{\text{deg}}$:
$$
D_\theta^{\mathrm{CDG}}(\boldsymbol{x}_\sigma;\sigma,\boldsymbol{c})=
D_\theta(\boldsymbol{x}_\sigma;\sigma,\boldsymbol{c})
+(w-1)\big(D_\theta(\boldsymbol{x}_\sigma;\sigma,\boldsymbol{c})-D_\theta(\boldsymbol{x}_\sigma;\sigma,\boldsymbol{c}_{\text{deg}})\big).
$$
The central claim is geometric: the null prompt induces an entangled guidance signal because $\epsilon_\theta(x_t,c)-\epsilon_\theta(x_t,\emptyset)$ mixes content correction with global style and structure. CDG reframes the contrast from “good vs. null” to “good vs. almost good,” so that common denoising components cancel more effectively.

The degraded condition is built by partitioning text-encoder tokens into content tokens and context-aggregating tokens. A single knob $R_{\mathrm{deg}}\in[0,2]$ determines per-type degradation ratios, and a binary mask $\boldsymbol{m}$ is applied through
$$
\boldsymbol{c}_{\text{deg}}=\boldsymbol{m}\odot \boldsymbol{c}+(1-\boldsymbol{m})\odot \emptyset.
$$
Token importance is obtained from self-attention maps using Weighted PageRank,
$$
\boldsymbol{s}^{(k+1)}=\frac{A^\top \boldsymbol{s}^{(k)}}{\|A^\top \boldsymbol{s}^{(k)}\|_1},
$$
rather than cross-attention summation, which the paper reports as misleading because it overemphasizes padding tokens. The default recommendation is $R_{\mathrm{deg}}=1.0$, termed the “semantic boundary,” which fully removes content tokens while preserving context tokens and bypasses WPR entirely.

The method is training-free, applies identically under DDPM, DDIM, and probability-flow ODE/SDE samplers, and is evaluated on Stable Diffusion 3, Stable Diffusion 3.5, FLUX.1-dev, and Qwen-Image. On MS-COCO 2017 val, SD3 improves from FID $35.69$ to $34.05$, CLIPScore from $31.73$ to $32.00$, Aesthetic Score from $5.66$ to $5.70$, and VQA Score from $91.44$ to $92.40$; SD3.5 improves from FID $34.56$ to $33.07$ and VQA from $91.94$ to $92.61$; FLUX.1-dev improves from FID $38.55$ to $37.11$; and Qwen-Image improves from FID $42.45$ to $39.02$ and VQA from $93.66$ to $93.93$. On GenAI-Bench, gains are reported for spatial relation, comparison, differentiation, and universal compositional reasoning. One-time WPR at the first denoising step adds only $+3.6\%$ wall time on SD3, while per-step recomputation adds $+47.2\%$ and is not recommended.

Within the diffusion literature, CDG is best understood as a negative-sample construction principle. It does not alter the scheduler and does not require external models or retraining; it changes the geometry of the guidance contrast.

## 4. CDG for synthetic asset-condition generation in power systems

In power-system reliability assessment, CDG is a framework for generating numerical and non-numerical in-group asset condition data when inspection records are unavailable or incomplete [2011.12352]. It integrates age-driven degradation modeling, condition correlation modeling, categorical distribution modeling, probabilistic diversification, and incorporation of expert knowledge. The target setting is ordinary in-group assets such as cables, conductors, poles, towers, service transformers, and switchgears, which are typically inspected periodically rather than monitored continuously.

Four degradation models are used for numerical, non-destructive conditions:
$$
C(t)=a\,t+b,\qquad
C(t)=b\,e^{at},\qquad
C(t)=a\,\ln(t)+b,\qquad
C(t)=b\,t^a.
$$
CDG further allows linear combinations of degradation forms,
$$
C_{\text{deg, combo}}(t)=\sum_{n=1}^{P}\alpha_n C_n^{(\text{deg})}(t),
$$
and combines degradation and correlation terms through
$$
C_{\text{combo}}(t)=\sum_{n=1}^{P}\alpha_n C_n^{(\text{deg})}(t)+\sum_{m=1}^{Q}\gamma_m C_m^{(\text{corr})}(t).
$$
Correlation is modeled by second-degree polynomial regression using previous inspection attributes. For numerical diversification, the generated value at age $x$ is treated probabilistically:
$$
C_x\sim \mathcal{N}(\mu_x,\sigma_x^2),\qquad \mu_x\equiv C_{\text{combo}}(x).
$$
For non-numerical conditions, the paper either maps ratings to numeric values or models them directly by age-dependent categorical probabilities $P(C=i\mid \mathrm{Age}=x)=\pi_i(x)$.

Validation uses two public datasets: 45-ft wood poles with inspection records from 1998, 2008, and 2018, and 20 kV XLPE cable segments with inspection records from 2010, 2013, 2016, and 2019. When both age and last inspection conditions are used, examples for cables include PD with KL divergence $0.011$ versus benchmark $0.193$ and MAPE $2.75\%$, TDS with KL $0.003$ versus $0.065$ and MAPE $3.26\%$, and DTD with KL $0.010$ versus $0.135$ and MAPE $3.22\%$. For poles, shell-thickness and circumference attributes achieve MAPE values around $3.81\%$ to $7.07\%$, while categorical mismatch percentages are approximately $3.4\%$, $3.7\%$, and $2.1\%$ for visual condition, surface condition, and woodpecker hole. When correlation is disabled and age alone is used, MAPE increases while KL divergence remains low relative to the benchmark. Holistic health-index prediction from generated conditions yields cable HI MAPE of approximately $2.1\%$ and pole HI mismatch percentage of approximately $1.2\%$.

The framework is then coupled to reliability analysis on an IEEE distribution feeder with 128 nodes and 1200 cable segments. Sequential Monte Carlo Simulation under a “run-to-failure” scenario is used with total ownership cost
$$
\mathrm{TOC}=\mathrm{PRC}+\mathrm{RRC}+\mathrm{FC},
$$
where failure cost is computed from Value of Lost Energy. In the replacement-strategy study, replacing approximately $70$ unhealthy cables per year minimizes TOC at approximately $\$702{,}580$ over 2019–2028.

Here CDG is generative and asset-centric. Its output is synthetic condition data with uncertainty quantification, intended to enable downstream reliability risk quantification and asset-management optimization.

## 5. CDG under partial observability: constrained IOHMMs and POMDP control

A fourth instantiation uses CDG as an inferred concept for operating-condition optimization guided by real-time degradation signals from multiple sensors [2512.06682]. The framework formulates a POMDP $(S,A,T,R,\Omega,O,\gamma)$ in which the hidden state space is $S=\{1,2,\ldots,K,F\}$, with left-to-right monotone degradation, absorbing failure, and preventive maintenance modeled as a reset to state $1$.

The observation model begins with a constrained Input-Output Hidden Markov Model. Transitions are action-dependent:
$$
[\mathbf{A}^{(a)}]_{k\ell}=P(S_t=k\mid S_{t-1}=\ell, A_t=a),
$$
with $[\mathbf{A}^{(a)}]_{k\ell}=0$ for $k<\ell$. Continuous multi-sensor observations are discretized through a Gaussian Mixture Model:
$$
P(o_i\mid \mathbf{O})=
\frac{\phi_i\,\mathcal{N}(\mathbf{O}\mid \mu_i,\sigma_i)}
{\sum_{j=1}^{J}\phi_j\,\mathcal{N}(\mathbf{O}\mid \mu_j,\sigma_j)}.
$$
Belief updates follow Bayesian filtering:
$$
b_o^a(s')=
\frac{O(o\mid s',a)\sum_{s\in\mathcal{S}}T(s,a,s')\,b(s)}
{\sum_{s'\in\mathcal{S}}O(o\mid s',a)\sum_{s\in\mathcal{S}}T(s,a,s')\,b(s)},
$$
and action selection is given by the infinite-horizon discounted Bellman equation
$$
V^*(b)=\max_{a\in A}\left[R(b,a)+\gamma\sum_{o\in\Omega}\Pr(o\mid a,b)\,V^*(b_o^a)\right].
$$
Parameter estimation is performed by generalized EM, and the POMDP is solved by Point-Based Value Iteration.

Two case studies are reported. For XJTU-SY bearing degradation data, eleven time-domain vibration features are extracted and six latent states are learned, with state $6$ corresponding to failure. Using 500 simulations of 50,000 cycles each, the proposed POMDP approach yields cumulative reward mean $14617.09$ with standard deviation $630.71$, compared with $11127.31\pm653.41$ for fixed capacity C1, $5846.04\pm616.38$ for C2, and $-9213.23\pm774.27$ for C3. For NASA C-MAPSS FD001 turbofan engines, the learned transition matrix has a left-to-right structure, predicted RUL tracks ground truth with credible bands that widen near end of life, and preventive maintenance dominates operation when posterior mass concentrates on the high-risk state. Hidden-state selection via AIC/BIC favors $K=6$, with AIC $1920.04$ and BIC $2368.05$.

This version of CDG is sequential and partially observable. Unlike the robust O\&M formulation, which embeds degradation directly in deterministic or robust constraints, the POMDP formulation acts on beliefs over latent health states.

## 6. Common themes, distinctions, and limitations

Across these papers, CDG consistently couples degradation information to action selection, but the nature of both “condition” and “guidance” varies sharply. In robust O\&M, condition signals parameterize uncertainty sets and degradation accumulation constraints [2311.10966]. In diffusion guidance, the degraded condition is a deliberately weakened semantic prompt used to generate a cleaner contrastive direction [2603.10780]. In power-system reliability, CDG synthesizes missing condition records from interpretable degradation and correlation models [2011.12352]. In the POMDP framework, multi-sensor observations are transformed into beliefs over latent degradation states that guide capacity and maintenance decisions [2512.06682].

Several distinctions follow. First, only the diffusion and power-system papers explicitly name their frameworks CDG; in the other two, the label is an interpretive overlay. Second, the term “degradation” is domain-specific: it refers to physical wear in asset-management papers and to selective semantic erasure in the diffusion paper. Third, “guidance” may mean robust prescriptive optimization, synthetic-data generation, or sampling-direction modification. A plausible implication is that the shared terminology reflects an abstract design pattern—construct a degraded or condition-derived reference, then use its contrast with a target signal to improve decisions—rather than disciplinary convergence on a common mathematical object.

The limitations are correspondingly domain-specific. The multi-asset O\&M formulation assumes linear decision-dependence in the optimization embedding, full restoration after maintenance or failure, Brownian residual stochasticity, and budgeted uncertainty sets without parametric distribution assumptions. The diffusion formulation can underperform when context-token degradation is too aggressive, when rare vocabulary is misidentified by the importance heuristic, or when prompt length and tokenization complicate $\lambda_{\text{block}}$ and ratio tuning. The power-system synthesis framework depends on the representativeness of historical or expert knowledge, uses second-degree polynomial correlation, and does not explicitly model spatial or environmental covariance. The POMDP framework assumes discrete-time, discrete-state Markov degradation, left-to-right monotonicity, observation discretization by GMM, and offline PBVI computation with online belief updates.

Taken together, the literature shows that CDG is best treated as a cross-domain label for methods that turn degradation-aware information into operational leverage. The concept is stable at that level of abstraction, but its concrete implementations belong to distinct technical traditions: robust optimization, diffusion guidance, synthetic condition-data generation, and partially observable sequential decision-making.

Source: https://www.emergentmind.com/topics/condition-degradation-guidance-cdg