MRC: Multi-Domain Applications in Research

• MRC is a multifaceted term, commonly representing machine reading comprehension in NLP, which underpins extractive QA and structured prediction tasks.
• In wireless communications, MRC (maximum ratio combining) optimizes signal reception by weighting channels based on quality, enhancing massive MIMO and NOMA systems.
• MRC also denotes specialized frameworks like the Market Regime Council for portfolio management and the Molonglo Reference Catalogue in radio astronomy.

In the cited arXiv literature, MRC denotes several distinct technical objects rather than a single unified concept. The most common usage is machine reading comprehension, a family of NLP tasks in which a model answers natural-language questions from a passage, often by extracting a span, selecting an option, or deciding that no answer exists (Bian et al., 2021, Putri et al., 2022, Ma et al., 10 Mar 2025). In communications and signal processing, MRC denotes maximum ratio combining or maximal-ratio combining, a receiver strategy studied in massive MIMO, cooperative NOMA, and cell-free architectures (Ganji et al., 2017, Liang et al., 2014, Yuan et al., 2020). More specialized usages include Market Regime Council, a cooperative multi-agent LLM system for online portfolio management (Pei et al., 23 May 2026), and the Molonglo Reference Catalogue, a radio-source catalogue used in H I absorption surveys (Aditya et al., 2023).

1. Principal expansions and disciplinary usage

The abbreviation appears with materially different semantics across research areas. The following map captures the usages explicitly documented in the cited papers.

Expansion of MRC	Domain	Representative use
Machine Reading Comprehension	NLP / NLU	extractive QA, multiple-choice QA, unanswerable QA
Maximum Ratio Combining / maximal-ratio combining	wireless communications	massive MIMO, NOMA, radio-stripe fronthaul
Market Regime Council	multi-agent LLM decision systems	Shapley-based portfolio allocation
Molonglo Reference Catalogue	radio astronomy	FLASH pilot survey target selection

Within NLP, machine reading comprehension has also become a modeling substrate for tasks not usually framed as question answering. The cited literature reformulates named-entity recognition, semantic role labeling, and dependency parsing as MRC-style prediction problems, suggesting that “MRC” in this domain often denotes both a benchmark family and a general interface for conditioning structured prediction on natural-language queries (Zhang et al., 2023, Wang et al., 2021, Gan et al., 2021).

2. Machine reading comprehension in natural language processing

In its extractive form, machine reading comprehension is the task of answering natural-language questions by locating an appropriate span in a document. Formally, given a document $\mathcal{D}$ with tokens $\{w_1,\dots,w_L\}$ and a question $Q$, an extractive MRC model selects a contiguous answer span $a = w_{i:j} = (w_i,\dots,w_j)$ in $\mathcal{D}$ (Bian et al., 2021). This formulation underlies standard benchmark evaluations such as SQuAD 1.1, NewsQA, TriviaQA, and SimpleNQ, typically scored with Exact Match and token-level F1 (Bian et al., 2021).

A major line of work concerns unanswerability. The Indonesian benchmark IDK-MRC was created because existing Indonesian MRC datasets lacked unanswerable questions. It combines automatic and manual unanswerable-question generation and, together with answerable questions, consists of more than 10K questions in total (Putri et al., 2022). Its train/dev/test splits contain 5,042 / 382 / 422 answerable and 4,290 / 382 / 422 unanswerable questions, respectively, and its unanswerable set is balanced across six types: negation, antonym, entity swap, question-tag swap, specific condition, and other (Putri et al., 2022). On the reported XLM-R comparison, training on IDK-MRC yields UnAns EM = 88.29, UnAns F1 = 88.29, Overall EM = 74.86, and Overall F1 = 81.37, versus 66.78 / 66.78 / 59.00 / 65.89 for TranslatedSQuAD and 0.90 / 0.90 / 33.01 / 39.74 for TyDiQA (Putri et al., 2022).

A second line concerns benchmark completeness. MRCEval introduces a three-level taxonomy of 13 distinct RC skills: six under context comprehension, three under external knowledge comprehension, and four under reasoning (Ma et al., 10 Mar 2025). It contains 2,103 multi-choice questions and evaluates 28 open-source and proprietary models with accuracy as the sole metric (Ma et al., 10 Mar 2025). The reported leaders are Qwen-max-2025 (59.4%), Gemini-2.0-flash (59.3%), o3-mini (59.0%), and DeepSeek-R1-Distill-Qwen-14B (56.4%), while counterfactual faithfulness remains a universal blind spot (Ma et al., 10 Mar 2025). This benchmark positions MRC not merely as span extraction but as a broader operationalization of reading comprehension.

3. MRC as a reformulation framework for structured prediction

Several papers use machine reading comprehension as a conversion layer for tasks that are not originally QA. In NER-to-MRC, Zhang et al. completely frame named-entity recognition as a machine reading comprehension problem (Zhang et al., 2023). The passage is the original token sequence $P = X$, the question is the fixed universal prompt “What kind of entity is this?”, and each option $O_j$ is the natural-language description of an entity type taken from the annotation guidelines (Zhang et al., 2023). The model input for option $j$ is the concatenation $[P;Q;O_j]$, processed in parallel across options with a shared DeBERTa-v3 encoder, followed by HRCA layers that review the question, read the option with cross-attention to the question, and find the passage with cross-attention to the enriched option representation (Zhang et al., 2023). The prediction tensor is $M^{pred}\in\mathbb{R}^{k\times n_0\times 2}$, trained with per-token, per-option binary cross-entropy:

$\{w_1,\dots,w_L\}$0

with overall loss

$\{w_1,\dots,w_L\}$1

Across six benchmark datasets from three domains, the system reports state-of-the-art performance without external data, including 65.61 F1 on WNUT-16, where the paper states an improvement of +11.24% over the best prior model (Zhang et al., 2023).

In semantic role labeling, Wang et al. formalize predicate disambiguation as multiple-choice MRC and argument labeling as extractive MRC (Wang et al., 2021). Candidate predicate senses are retrieved from PropBank frame files, each paired with its sense description $\{w_1,\dots,w_L\}$2, and the input is

$\{w_1,\dots,w_L\}$3

A sigmoid-scored FFN predicts the correct sense, while argument labeling uses role-specific queries such as “What are the X arguments of predicate Y with meaning Z?” for core roles and “What are the W modifiers of predicate Y?” for non-core roles (Wang et al., 2021). RoBERTa-based results include 97.3 predicate-disambiguation accuracy on CoNLL-2009 WSJ and 93.3 / 87.2 F1 on CoNLL-2009 WSJ/Brown for dependency SRL, with 90.0 / 85.1 / 88.3 F1 on CoNLL-2005 WSJ, CoNLL-2005 Brown, and CoNLL-2012 for span SRL (Wang et al., 2021).

In dependency parsing, Gan et al. model parsing as MRC-based span-span prediction (Gan et al., 2021). Candidate subtrees are proposed as spans $\{w_1,\dots,w_L\}$4, and span linking is formalized as MRC where the context is the full sentence, the query is a marked child span, and the answer is the parent span (Gan et al., 2021). The query is encoded by inserting the markers <sos>, <sor>, <eor>, <eos> around the child span in the sentence. Separate scores are predicted for the parent head, start, end, and dependency label, and a reverse child-query can be added for mutual consistency (Gan et al., 2021). On PTB, the RoBERTa-based parser reaches 97.24% UAS / 95.49% LAS; on CTB it reaches 93.14% UAS; and it wins 11 of 12 UD languages with average +0.30 LAS (Gan et al., 2021).

4. Learning regimes for MRC: self-supervision, masking, and domain adaptation

One research direction seeks to align pre-training more closely with the span-spotting structure of MRC. In “Bridging the Gap between LLM and Reading Comprehension: Unsupervised MRC via Self-Supervision”, the pretext task Spotting-MLM masks an entire occurrence of an informative repeated span and requires the model to predict the start and end indices of another occurrence (Bian et al., 2021). The masked query representation $\{w_1,\dots,w_L\}$5 is obtained from the [MASK] token, and start and end probabilities are computed with bilinear scoring:

$\{w_1,\dots,w_L\}$6

The loss is the sum of negative log-likelihoods over gold start and end positions (Bian et al., 2021). On four unsupervised benchmarks, the reported average score is 32.0 EM / 42.6 F1, compared with 29.6 / 39.1 for Fabbri et al. (2020), and on SQuAD-Adv “AddSent” the model reaches 39.9 F1 (Bian et al., 2021).

A second direction studies whether masking length distribution should match downstream answer-length distribution. The Chinese study “Analysing the Effect of Masking Length Distribution of MLM” defines four MRC tasks—short-span extraction, long-span extraction, short multiple-choice cloze, and long multiple-choice cloze—and creates four corresponding Chinese datasets (Zeng et al., 2021). It then trains four MLMs whose masking-span-length distributions are set to

$\{w_1,\dots,w_L\}$7

where $\{w_1,\dots,w_L\}$8 is the number of answers of length $\{w_1,\dots,w_L\}$9 in dataset $Q$0 (Zeng et al., 2021). In every reported case, the MLM whose mask-length distribution matches the target answer distribution achieves the highest downstream score, although the gains are described as modest—approximately 1–2% absolute—and human performance remains far higher (Zeng et al., 2021). This suggests that MRC performance is sensitive not only to architecture but also to the structural statistics of pre-training corruption.

A third direction addresses domain shift. BioADAPT-MRC adapts from general-purpose QA to biomedical QA without using synthetic or human-annotated biomedical target data (Mahbub et al., 2022). The model comprises a shared BioELECTRA feature extractor $Q$1, an answer predictor $Q$2, and a domain similarity discriminator $Q$3 with a triplet-loss head and an auxiliary QA head (Mahbub et al., 2022). The combined objective is

$Q$4

implemented through a gradient-reversal layer (Mahbub et al., 2022). Using SQuAD-1.1 as source data and BioASQ-7b/8b/9b contexts as unlabeled targets, the model reports 0.45 / 0.64 / 0.53 on BioASQ-7b, 0.38 / 0.62 / 0.47 on BioASQ-8b, and 0.54 / 0.74 / 0.62 on BioASQ-9b for SAcc / LAcc / MRR, together with 0.21 EM / 0.40 F1 on emrQA (Mahbub et al., 2022).

5. Maximum ratio combining in wireless communications

In communications, MRC denotes maximum ratio combining or maximal-ratio combining, a receiver rule that weights branches by channel quality. In the dual-branch interference-aware model analyzed under spatial interference correlation and Nakagami fading, the post-combiner SINR is

$Q$5

where the two interference powers are correlated because they arise from the same Poisson field of interferers (Tanbourgi et al., 2013). The paper derives an exact success-probability expression for dual-branch MRC, shows that assuming independent interference across antennas is highly optimistic, and shows that the common full-correlation approximation is slightly pessimistic but becomes exact as the interferer Nakagami parameter $Q$6 (Tanbourgi et al., 2013).

In uplink massive MIMO, the cited stochastic-geometry analysis decomposes the interference under MRC into intra-cell interference, inter-cell interference, and pilot-contamination interference (Liang et al., 2014). For practical but still large numbers of BS antennas, the key observation is that intra-cell interference accounts for the dominant portion of the total interference under MRC reception, whereas ZF substantially reduces this component while leaving inter-cell interference and pilot contamination almost unchanged (Liang et al., 2014). The paper also notes that pilot-contamination interference and inter-cell interference have roughly equal mean powers, but pilot contamination has a much wider distribution range under strong shadowing (Liang et al., 2014).

Timing asynchrony fundamentally changes MRC behavior in large-scale MU-MIMO. In “On the Performance of MRC Receiver with Unknown Timing Mismatch”, neglecting unknown user delays causes the MRC achievable rate to saturate because residual ISI persists even as $Q$7 grows (Ganji et al., 2017). Under perfect CSI and power scaling $Q$8, ordinary MRC converges to

$Q$9

which is capped at high $a = w_{i:j} = (w_i,\dots,w_j)$0 by average ISI (Ganji et al., 2017). The proposed MRC-ZF post-processing removes average ISI using only the distribution of the timing mismatch, and the paper reports restored large-$a = w_{i:j} = (w_i,\dots,w_j)$1 power scaling together with close agreement between theory and simulation (Ganji et al., 2017).

In vehicular communications at road intersections, cooperative NOMA with MRC yields closed-form outage expressions and improves over both standard cooperative NOMA and MRC-OMA (Belmekki et al., 2020). Under the reported settings, MRC-NOMA reduces the outage of $a = w_{i:j} = (w_i,\dots,w_j)$2 by about 34% versus standard NOMA relay and exceeds 60% outage gain over MRC-OMA for $a = w_{i:j} = (w_i,\dots,w_j)$3 bit/s/Hz (Belmekki et al., 2020). The paper also shows that outage peaks near the road intersection, where both perpendicular roads contribute interference (Belmekki et al., 2020).

In cell-free massive MIMO with radio stripes, MRC also appears in the fronthaul design. Each AP forwards only the local matched-filter outputs, giving a fronthaul load of

$a = w_{i:j} = (w_i,\dots,w_j)$4

real scalars per coherence (Yuan et al., 2020). The proposed Q-LMMSE reconstructs interference structure from the covariance of the MRC stream via a small $a = w_{i:j} = (w_i,\dots,w_j)$5 SVD and computes

$a = w_{i:j} = (w_i,\dots,w_j)$6

The paper reports that Q-LMMSE increases spectral efficiency by 347.1% over MRC and 51.3% over N-LMMSE in the correlated-channel case with $a = w_{i:j} = (w_i,\dots,w_j)$7, while retaining the same MRC fronthaul (Yuan et al., 2020).

6. Other specialized meanings: Market Regime Council and the Molonglo Reference Catalogue

In multi-agent LLM decision systems, MRC denotes Market Regime Council (Pei et al., 23 May 2026). This framework treats specialist agents as players in a transferable-utility game, evaluates all non-empty coalitions, computes exact Shapley values, and uses those credits for online weighting (Pei et al., 23 May 2026). With $a = w_{i:j} = (w_i,\dots,w_j)$8 agents, it runs three deliberation stages: private single-agent portfolios, pairwise debate portfolios, and a grand-coalition portfolio with a readout operator (Pei et al., 23 May 2026). Coalition value is defined from exponentially weighted estimates of Sharpe and annualized return,

$a = w_{i:j} = (w_i,\dots,w_j)$9

and each period also applies a Bayesian adaptive mixture, regime-dependent multipliers, and a five-layer causal trace (Pei et al., 23 May 2026). Over 1,037 trading days, 13 crypto assets, and five seeds, the reported performance is CR = 440.1%, SR = 1.51, MDD = 34.1%, and IR = 0.47, ranking first on CR, SR, and IR among active baselines and achieving the lowest MDD among active methods (Pei et al., 23 May 2026).

In radio astronomy, MRC denotes the Molonglo Reference Catalogue. The FLASH pilot survey searched for associated H I 21-cm absorption against bright radio sources from the MRC 1-Jy sample (Aditya et al., 2023). The final sample contained 62 MRC galaxies and quasars with $\mathcal{D}$0, yielding three new detections of associated H I 21-cm absorption (Aditya et al., 2023). After smoothing non-detections to 100 km s$\mathcal{D}$1 and excluding five shallow spectra, the overall detection fraction is reported as 1.8%$\mathcal{D}$2, or 1/57 = 1.8% (95% confidence interval 0.3–5.8%) (Aditya et al., 2023). All three detections are in the peaked-spectrum or compact steep-spectrum class, for which the detection fraction is 23%$\mathcal{D}$3 (Aditya et al., 2023). The strongest system, MRC 0531–237, has integrated optical depth $\mathcal{D}$4, described as the strongest found to date in the sample (Aditya et al., 2023).

Taken together, the cited literature shows that MRC is not a stable single-term entry across the arXiv ecosystem. In NLP it is primarily a question-answering and task-reformulation framework; in wireless communications it is a receiver-combining principle with deep links to interference structure, synchronization, and fronthaul design; and in more specialized contexts it names a coalition-based financial decision system and a long-standing radio-source catalogue (Ma et al., 10 Mar 2025, Ganji et al., 2017, Pei et al., 23 May 2026, Aditya et al., 2023). A plausible implication is that any technical use of the abbreviation requires immediate domain disambiguation.