# q-fin.ST Published articles for q-fin.ST. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Isotonic Quantile Regression Averaging for uncertainty quantification of electricity price forecasts DevFeed: [Isotonic Quantile Regression Averaging for uncertainty quantification of electricity price forecasts]() Original publisher: [Read original article]() Author: Arkadiusz Lipiecki, Bartosz Uniejewski Published: 2026-11-09T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Machine learning](), [linear-regression](), [IoT sensor data forecasting]() Tags: [battery](), [complexity](), [computational-complexity](), [conformal-prediction](), [coverage](), [cs-lg](), [electricity](), [energy](), [forecast](), [forecasting](), [markets](), [models](), [order](), [practical](), [prediction](), [probabilistic](), [q-fin-st](), [quantile](), [regression](), [reliability](), [stat-ap](), [uncertainty](), [uncertainty-quantification]() ### AI overview Isotonic Quantile Regression Averaging (iQRA) adds stochastic order constraints to Quantile Regression Averaging (QRA) to produce probabilistic forecasts from ensembles of point forecasts. The authors present the constraints as hyperparameter-free regularization that lowers computational complexity while maintaining competitive accuracy, and apply the same principle to smoothed QRA. In a study of two European day-ahead electricity markets, iQRA and smoothed iQRA rank among leading postprocessing methods for reliability and sharpness; the abstract reports well-calibrated intervals and stronger reliability for smoothed iQRA than coverage-oriented conformal prediction. A battery-arbitrage evaluation finds that leading probabilistic methods yield the highest profit per trade and that probabilistic risk control improves trading value. These results are reported in the abstract; it provides no detailed methodology or numerical comparisons. ### Source excerpt arXiv:2507.15079v2 Announce Type: replace Abstract: Quantifying the uncertainty of forecasting models is essential to assess and mitigate the risks associated with data-driven decisions, especially in volatile domains such as electricity markets. In this paper, we propose a computationally efficient isotonic variant of Quantile Regression Averaging (QRA), called Isotonic Quantile Regression Averaging (iQRA), for generating probabilistic forecasts from ensembles of point forecasts. Building on QRA, we introduce stochastic order constraints as a hyperparameter-free regularization mechanism that reduces computational complexity while maintaining competitive forecast accuracy. We further demonstrate that the same isotonic regularization principle can be applied to the smoothed estimator of QRA. In an extensive forecasting study of two European day-ahead electricity markets, we show that iQRA and its smoothed variant consistently rank among the top state-of-the-art postprocessing methods in terms of reliability and sharpness. It produces well-calibrated prediction intervals, with smoothed iQRA providing superior reliability even when compared against coverage-oriented conformal prediction. We further introduce a battery arbitrage evaluation, showing that the leading probabilistic methods achieve the highest profit per trade and that probabilistic risk control improves the economic value of trading decisions. Isotonic regularization offers a scalable hyperparameter-free variable selection for linear quantile regression, making iQRA attractive for practical decision-support systems in energy markets. ## Decomposing Rank Reversals under Temporal Nonstationarity: Alignment, Aggressiveness, and Instability DevFeed: [Decomposing Rank Reversals under Temporal Nonstationarity: Alignment, Aggressiveness, and Instability]() Original publisher: [Read original article]() Author: Deying Song, Xueying Ding Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [model architecture]() Tags: [adaptive](), [alignment](), [architectures](), [benchmarking](), [complexity](), [cs-lg](), [evaluation](), [financial](), [linear](), [memory](), [model](), [model-agnostic](), [neural](), [performance](), [q-fin-st](), [r](), [rank](), [rankings](), [regression](), [reverse](), [shift](), [temporal]() ### AI overview The paper attributes model rank reversals across training horizons under temporal nonstationarity to predictable surges in miscalibrated aggressiveness. It decomposes out-of-sample R² into target alignment, miscalibrated aggressiveness, and predictive instability, then derives complexity phase boundaries and closed-form optimal-memory equations for ridge regression under local drift. Surrogate bounds extend the theoretical guarantees to more complex architectures. Benchmarks of 14 linear, tree-based, and neural models on large financial panels are reported as validating these dynamics. The supplied text is an abstract, so it does not provide details of the equations, benchmark design, or quantitative results. ### Source excerpt arXiv:2611.04116v1 Announce Type: new Abstract: Under temporal nonstationarity, model performance rankings reverse across training horizons due to heterogeneous sensitivities to memory. While existing literature addresses adaptive lookbacks, effective model complexity, and empirical forecast evaluation in isolation, no unified framework explains the structural mechanics driving these rank reversals. We bridge this gap by deriving a model-agnostic decomposition of out-of-sample $R^2$ into target alignment, miscalibrated aggressiveness, and predictive instability. Within this framework, we mathematically characterize complexity phase boundaries, establish closed-form optimal-memory equations for ridge regression under local drift, and extend these guarantees to complex model architectures via surrogate bounds. We validate these theoretical dynamics by benchmarking 14 linear, tree-based, and neural architectures across large-scale financial panels. In doing so, our framework demonstrates that empirical model rank reversals are driven by predictable surges in miscalibrated aggressiveness, providing a rigorous mechanistic diagnosis of generalization under temporal shift. ## Honest Small Decision Models for Finance: Scale, Specialization Collapse, and Mixed-Corpus Remediation DevFeed: [Honest Small Decision Models for Finance: Scale, Specialization Collapse, and Mixed-Corpus Remediation]() Original publisher: [Read original article]() Author: Ajitabh Kumar Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [AI Models](), [Fine-tuning](), [Finance]() Tags: [cs-lg](), [decision-models](), [finance](), [models](), [q-fin-st](), [remediation](), [scale](), [small](), [specialization]() ### AI overview The evaluated finance outcome models showed no detectable conditional edge: across sizes from 0.8B to 32B parameters, accuracy matched the horizon-conditioned prior at 0.37-0.38, with calibrated probabilities tracking per-horizon base rates. Calibration was strongest at 2B (ECE 0.026) and worsened to 0.043-0.044 at 9B-32B under prior drift. Narrow fine-tuning reduced general decision competence by 9-45 percentage points, with the loss shrinking at larger scales. A mixed-corpus adapter retained policy-task performance and near-specialist outcome calibration while matching or slightly exceeding the generalist on the public benchmark. The authors describe dispatchable confidence bands for rule-determined answers and review routing for uncertain real-world outcomes. ### Source excerpt arXiv:2611.03755v1 Announce Type: new Abstract: Financial workflows need fast, calibrated, auditable decisions, not just generated text. We train System-1 decision models, small language models fine-tuned to answer typed questions (choice/score/noul) with a probability in a single forward pass, across two model families (K2-Horizon and Qwen3.5/3.8) and nine rungs from 0.8B to 32B parameters, using one frozen LoRA recipe on a public decision benchmark and two new finance corpora: a programmatic-gold policy corpus (rule-reading with exact labels) and a reality-labeled outcome corpus (19,640 realized forward-return questions in walk-forward splits, plus a sealed 2026 test of 1,360 questions). Narrow fine-tuning costs general decision competence 9-45 percentage points, a data-mixture effect that shrinks with model scale and is associated with narrow-gradient volume. On reality-labeled questions, every trained rung from 0.8B to 32B lands at the horizon-conditioned prior baseline (accuracy 0.37-0.38, no detectable conditional edge under our controls); the models respond with flat, well-calibrated probabilities that track per-horizon base rates: naturally honest uncertainty, with a mid-scale calibration sweet spot (ECE 0.026 at 2B, degrading to 0.043-0.044 at 9-32B under prior drift). A mixed-corpus remediation carries all competencies in one adapter at modest cost: it matches or slightly exceeds the generalist specialist on the public benchmark while holding the policy ceiling and near-specialist outcome calibration, yielding a deployable artifact with per-domain temperature regimes and a band-dispatch escalation analysis: sharp, dispatchable bands where rules determine answers; flat, review-routing bands where reality does not. ## Hardware-Aware Design and Layout Selection in Quantum Reservoir Computing for Insurance Claims Forecasting DevFeed: [Hardware-Aware Design and Layout Selection in Quantum Reservoir Computing for Insurance Claims Forecasting]() Original publisher: [Read original article]() Author: Ren\\'ata Rusn\\'akov\\'a, Eugen \\v{S}lapak, Gaurav K. Pandey, Juraj Gazda Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Quantum Computing](), [quantum mechanics]() Tags: [computing](), [cs-lg](), [design](), [forecasting](), [hardware](), [insurance](), [layout](), [q-fin-st](), [quant-ph](), [quantum](), [reservoir-computing](), [selection]() ### AI overview A six-qubit quantum reservoir achieved next-day motor-insurance claim forecast RMSEs of 5.66-6.12 claims per day on 232 held-out days, versus 5.36 for ridge regression; four candidates were not significantly different from ridge under a 95% block bootstrap. On IBM Kingston, hardware-aware layouts had lower RMSE than Qiskit optimization level 1 in all 21 paired rounds and lower maximum transfer distortion (7.59 versus 11.84 claims per day), while avoiding severe failures observed with some random-native layouts. A brief probe using training inputs detected systematic forecast shifts before test runs. Longer replay windows increased circuit cost without consistent accuracy or transfer gains. These findings support competitive performance and more consistent hardware transfer for this task. ### Source excerpt arXiv:2611.03891v1 Announce Type: cross Abstract: Quantum reservoir computing (QRC) processes temporal information through quantum dynamics, but it remains unclear whether small QRC models can forecast real data competitively with classical models and whether their predictions survive the transfer from ideal simulation to noisy hardware. We address both questions for a six-qubit gate-based reservoir that forecasts next-day motor-insurance claim counts. Five reservoir candidates were selected and frozen using hardware-native interaction graphs, Trotterized Ising dynamics, memory and measurement-robustness criteria, circuit cost, and calibrated-noise screening. On 232 held-out test days, the ideal candidates reached root-mean-square errors (RMSEs) of 5.66-6.12 claims/day, compared with 5.36 for ridge regression. For four of the five, the difference from ridge was not significant under a 95% block bootstrap, and their accuracy was comparable with classical recurrent and reservoir baselines. On the 156-qubit IBM Kingston processor, we ran 63 evaluations in 21 paired rounds, comparing hardware-aware, random-native, and Qiskit optimization level 1 layouts. Hardware-aware placement gave lower RMSE than Qiskit in every round, with a maximum transfer distortion of 7.59 claims/day (11.84 for Qiskit), and avoided the severe transfer failures seen for some random layouts. Most of the hardware distortion appeared as a systematic forecast shift, which a short hardware probe on training inputs detected before the test runs. Longer replay windows raised circuit cost without a consistent gain in forecasting accuracy or hardware transfer. Small QRC models can thus reach competitive accuracy on this task, and hardware-aware layout selection makes their transfer to IBM Kingston more consistent. ## Thinking Fast, Trading Late: System One Decision Models vs the Market DevFeed: [Thinking Fast, Trading Late: System One Decision Models vs the Market]() Original publisher: [Read original article]() Author: Amir Dehkordi (Independent Researcher), Alireza Kargarzadeh (TailState Intelligence Ltd), Nariman Khaledian (Independent Researcher), Navid Parvini (Zanista AI Ltd), Sid Ghatak (Increase Alpha, LLC), Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence](), [arXiv: Machine Learning]() Topics: [Artificial Intelligence](), [Data analysis](), [Finance]() Tags: [ai-decision-models](), [company-news](), [costs](), [cs-ai](), [cs-cl](), [cs-lg](), [deflated-sharpe](), [large-language-models-llms](), [masking](), [open-weights-model](), [protocol](), [q-fin-st](), [s-p-500](), [validation]() ### AI overview OLIVIA tests whether AI decision models can predict stock returns without look-ahead, selection bias or company-name memorization, using point-in-time masked news, validation-only selection, a hash-frozen analysis plan and a single logged unblinding of the test period. A fine-tuned open-weights model replicated its signal on two validation quarters, but its daily information coefficient fell to 0.020 (t = 1.7) in the blinded January-July 2026 test. Three commercial models tracked overnight moves to the open (IC 0.074-0.077), but not subsequent moves, and no preregistered strategy survived costs and a deflated Sharpe ratio. The models could not distinguish overreaction from underreaction; commercial models also identified masked companies in 62%-74% of cases versus a 28% baseline, showing that masking names did not conceal identities. ### Source excerpt arXiv:2611.03985v1 Announce Type: cross Abstract: Large language models (LLMs) read company news at scale, and backtests suggest that their readings predict stock returns. We ask how much of that predictability survives a test that rules out look-ahead, memorization and selection. We propose OLIVIA, a protocol for such tests: point-in-time inputs with company identities masked, selection on validation data only, an analysis plan frozen by cryptographic hash, and a test period unblinded once through a logged lock. We apply it to AI decision models, non-generative language models that return probabilities for typed questions, reading news about S&P 500 stocks: an open-weights model we fine-tuned on 2024-2025 data, three commercial models and standard baselines. The fine-tuned model's signal replicated on two validation quarters (daily information coefficients 0.055 and 0.065) but fell to 0.020 (t = 1.7) on the blinded January-July 2026 test. The commercial models' readings line up with the overnight move to the open (information coefficient 0.074 to 0.077) but not with the move after it, so the news is priced by the time a trader can act on it at the open. No model can tell over- from under-reaction, and no pre-registered strategy survives costs and a deflated Sharpe ratio. Commercial models identify masked companies in 62% to 74% of cases against a 28% baseline: masking hides names, not companies. The market reads the news first. ## Can We Stop the Backtest Yet? Anytime-Valid Promotion Tests for Adaptive Financial Research Agents DevFeed: [Can We Stop the Backtest Yet? Anytime-Valid Promotion Tests for Adaptive Financial Research Agents]() Original publisher: [Read original article]() Author: Eric Benhamou, Jean-Jacques Ohana, Beatrice Guez, Ethan Setrouk, Chamyl Saadi, Thomas Jacquot Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Finance](), [AI Chat](), [generative-ai-tools]() Tags: [abstract](), [adaptive](), [agents](), [ai](), [ai-tools](), [ai-use](), [alternatives](), [arxiv](), [audit](), [backtest](), [bounded](), [candidate](), [certificate](), [codex](), [compilation](), [component](), [cs-ai](), [financial](), [promotion](), [q-fin-st](), [research](), [tests]() ### AI overview PROMOTE is an anytime-valid protocol for testing adaptive trading-rule candidates: it assigns each candidate predictable testing wealth at birth and accumulates a cost-aware e-process using subsequent returns. When lifetime allocations total at most one, a weighted Ville argument bounds the chance of ever promoting a rule whose declared bounded net score has always had nonpositive conditional mean, even with dependent candidates, adaptive births, and indefinite monitoring. In a retrospective replay, repeated unadjusted tests crossed in 94.5% of 3,000 shared-sign null paths, compared with 0.0% for PROMOTE. With a 15-bp aligned ETF-component perturbation, PROMOTE had 70.1% target power and a median crossing session of 2,327. In a separate exact mean-zero experiment with agent-selected candidates and birth times, false-family crossings were 1.77% across 3,000 programs. The market replay yielded no promotions after costs. The certificate addresses historical false promotion under the declared null; it does not establish that an earlier edge remains profitable at deployment. ### Source excerpt arXiv:2611.03384v1 Announce Type: new Abstract: Financial research agents can generate and monitor trading rules faster than a human can document the resulting selection process. Candidates appear at data-dependent times, weak rules are abandoned, and promising ones are watched until they cross a reporting threshold. We introduce PROMOTE, a protocol that assigns every candidate predictable testing wealth at birth and accumulates a cost-aware e-process only from subsequent returns. If lifetime allocations sum to at most one, a weighted Ville argument bounds the probability of ever promoting a rule whose declared bounded net score has always had nonpositive conditional mean. The result permits dependent candidates, adaptive births, and indefinite monitoring. We audit the implementation in two settings. A retrospective pseudo-out-of-sample replay contains 24 panel--rule hypotheses, 3,000 shared-sign null paths, injected alternatives, and public ETF and industry data. Repeated unadjusted tests produce a family crossing in 94.5% of null paths, versus 0.0% for PROMOTE; at a 15-bp aligned ETF-component perturbation, PROMOTE has 70.1% target power and median crossing session 2,327. A separate exact mean-zero experiment lets an agent select candidate identities and birth times from past scores; the funded gate has 1.77% false-family crossings across 3,000 programs. The market replay produces zero promotions after costs. The certificate controls historical false promotion under the declared null; it is not a certificate that an earlier edge remains profitable at deployment. AI-use disclosure. Generative AI tools, including OpenAI Codex, assisted manuscript preparation. For this preprint version, Codex assisted with LaTeX editing, author and venue metadata, compilation, and layout checks. ## TensorLOB: Structure-Preserving Tensor Learning for Limit Order Book Forecasting DevFeed: [TensorLOB: Structure-Preserving Tensor Learning for Limit Order Book Forecasting]() Original publisher: [Read original article]() Author: Yuxuan Gu, Danilo Mandic Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Machine learning](), [Finance]() Tags: [benchmark](), [cs-lg](), [financial](), [forecasting](), [learning](), [limit-order-book](), [market-microstructure](), [q-fin-st](), [structure](), [tensors]() ### AI overview TensorLOB preserves the limit order book's time, market-side, price-level, and order-attribute dimensions by learning and predicting directly in the tensor domain. It extracts features through mode-wise tensor contractions and uses low-rank class-weight tensors for prediction. On the FI-2010 benchmark, it achieved the highest accuracy, recall, and F1 score among evaluated models at all four forecast horizons. The abstract presents this as a proof of concept; it provides no further detail here on performance margins or limitations. ### Source excerpt arXiv:2611.02866v1 Announce Type: new Abstract: The limit order book (LOB) records outstanding buy and sell orders across multiple price levels. Its state evolves over time across market side, price level, and order attribute, naturally forming a multiway structure. However, existing LOB models typically flatten these dimensions into a matrix model, thus obscuring the existing structural relationships. Starting from first principles and the market-microstructure intuition of the LOB, we design TensorLOB to operate entirely in the tensor domain. TensorLOB uses mode-wise tensor contractions for feature extraction and directly contracts the learned representation with low-rank class-weight tensors for prediction, retaining the distinct roles of time, side, level, and attribute modes. Experiments on the FI-2010 benchmark show that TensorLOB achieves the highest accuracy, recall, and F1 score across all four evaluated horizons. This proof of concept highlights the potential of high-dimensional modelling of multi-way latent information intrinsic to financial markets, which has so far been largely overlooked. ## A Controlled Comparison of Deep Learning Architectures for Multi-Horizon Financial Forecasting: Evidence from 918 Experiments DevFeed: [A Controlled Comparison of Deep Learning Architectures for Multi-Horizon Financial Forecasting: Evidence from 918 Experiments]() Original publisher: [Read original article]() Author: Nabeel Ahmad Saidd Published: 2026-11-02T05:00:00Z Content type: comparison Language: en Sources: [arXiv: Machine Learning]() Topics: [Predictive AI](), [Time Series](), [Hyperparameter optimization](), [Deep learning](), [Machine learning]() Tags: [architectures](), [bayesian](), [benchmark](), [comparison](), [cryptocurrency](), [cs-lg](), [deep-learning](), [error](), [evaluation](), [experiments](), [financial](), [financial-forecasting](), [forecasting](), [horizon](), [hyperparameter-optimization](), [linear](), [metric](), [prediction](), [q-fin-gn](), [q-fin-st](), [rank](), [ranking](), [time-series]() ### AI overview In a controlled benchmark of 918 runs across nine architectures, 12 assets, and 4- and 24-hour forecast horizons, ModernTCN ranked first overall (mean rank 1.333) and had the lowest RMSE at 18 of 24 asset-horizon points; PatchTST ranked second. The five-stage procedure used equal-budget Bayesian optimization, frozen category configurations, three-seed training, held-out aggregation, and rank-based validation. Ranking differences were statistically strong, while rankings remained consistent across horizons and errors rose at 24 hours. Architecture accounted for much more error variance than random seed under the reported analyses. Mean directional accuracy was 50.08%, so the results establish relative performance under these benchmark conditions, not strong directional prediction. ### Source excerpt arXiv:2603.16886v2 Announce Type: replace-cross Abstract: Reliable comparison of deep learning architectures for financial forecasting requires controlled evaluation that separates architectural effects from hyperparameter tuning, random initialization, forecast horizon, and asset characteristics. We present a five-stage benchmark using equal-budget Bayesian hyperparameter optimization, category-level configuration freezing, three-seed final training, held-out metric aggregation, and rank-based statistical validation. Nine architectures from Transformer, MLP/linear, convolutional, and recurrent families are evaluated with direct multi-step prediction from five OHLCV features across 12 assets (four cryptocurrency, four foreign exchange, four equity-index instruments), using chronological 70/15/15 splits of 30,000 hourly windows per asset at 4- and 24-hour horizons. The benchmark includes 918 runs (270 optimization trials, 648 final training runs). Across 24 asset-horizon points, ModernTCN has mean rank 1.333 and the lowest RMSE in 18 points (75.0%), followed by PatchTST (mean rank 2.000). Global ranking differences are strongly supported by an Iman-Davenport test (F(8,184)=101.36, p<1e-15), with 33 of 36 Holm-adjusted pairwise comparisons significant. Rankings are consistent across horizons for all 12 assets (rho>=0.683; Stouffer p=3.47e-10), while errors increase at the longer horizon. Architecture explains 99.90% of raw error variance and 68.33% after within-slot normalization excluding LSTM; seed variance contributes at most 0.04% (0.02% without LSTM). Mean directional accuracy is 50.08%. These findings provide a statistically validated comparison of nine architectures, identify ModernTCN and PatchTST as leading performers under the evaluated conditions, and offer a reproducible benchmark for multi-asset financial time-series forecasting. ## Sparse Autoencoders as Implicit Diversification Regularisers for Cross-Sectional Cryptocurrency Returns DevFeed: [Sparse Autoencoders as Implicit Diversification Regularisers for Cross-Sectional Cryptocurrency Returns]() Original publisher: [Read original article]() Author: Weijia Han, Lisha Qu Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Finance](), [Neural Network](), [Cryptocurrency]() Tags: [autoencoders](), [coding](), [cryptocurrency](), [cs-lg](), [factor](), [features](), [financial](), [portfolios](), [q-fin-pm](), [q-fin-st](), [return](), [risk]() ### AI overview The study evaluates sparse autoencoders (SAEs) as a way to represent cross-sectional cryptocurrency returns with a small number of active features, using equities as a secondary test and time-ordered comparisons with common benchmarks. SAE portfolios show consistently lower volatility, which the authors associate with lower correlations among selected assets. They do not outperform a simple contrarian strategy on risk-adjusted return, and the learned factor identities vary across training runs. The reported evidence supports a diversification effect, not claims of alpha or stable factors. ### Source excerpt arXiv:2610.39205v1 Announce Type: new Abstract: Factor models are central to asset pricing and portfolio construction, but neural factors are often difficult to interpret, and their effect on portfolio risk is unclear. A sparse autoencoder (SAE) may address both concerns by describing daily returns across many assets with only a few active features. We ask whether this sparse representation produces stable financial factors and better-diversified portfolios. We study cryptocurrency returns, use equities as a secondary test, and compare the SAE with common benchmarks in a time-ordered evaluation. We find consistently lower volatility in SAE portfolios and link this result to lower correlations among the selected assets. However, the SAE does not outperform a simple contrarian strategy on risk-adjusted return, and its factor identities are not stable across training runs. Our contribution is to show that sparse coding can support portfolio diversification, while clearly separating this risk effect from claims about alpha or stable factor identities. ## FinLLM-KG: Knowledge Graph-Enhanced LLMs for Financial Sentiment Analysis in Algorithmic Trading DevFeed: [FinLLM-KG: Knowledge Graph-Enhanced LLMs for Financial Sentiment Analysis in Algorithmic Trading]() Original publisher: [Read original article]() Author: Giorgos Iacovides, Oliver Hannibal, Wuyang Zhou, Danilo Mandic Published: 2026-10-28T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [база знаний](), [Large Language Model](), [Retrieval Augmented Generation (RAG)](), [Inference](), [Finance]() Tags: [algorithmic-trading](), [cs-cl](), [cs-lg](), [explainability](), [financial](), [knowledge-graph](), [llms](), [q-fin-st](), [q-fin-tr]() ### AI overview FinLLM-KG adds temporally grounded company knowledge from SEC 10-K filings to causal LLM sentiment predictions at inference time, without changing model weights. It supplies context as either a natural-language company profile or a small set of relevance-ranked factual triples, letting predictions draw on firm-level information as well as the article. The framework is described as model-agnostic, with almost zero computational cost, and links sentiment scores to sentences and retrieved facts to make their evidence inspectable for human review. Simulations reportedly show consistently better portfolio performance than the state-of-the-art baseline; the excerpt provides no quantitative results or evaluation details. ### Source excerpt arXiv:2610.36083v1 Announce Type: cross Abstract: Recent advances in Generative AI have substantially improved financial sentiment analysis through post-trained financial large language models (LLMs). However, existing approaches remain confined to predicting the sentiment of each article in isolation and operate as black boxes, which limits contextual understanding and hinders deployment in high-stakes financial decision-making. To this end, we introduce FinLLM-KG, a model-agnostic framework that combines causal LLMs with a financial knowledge graph, in order to inject temporally grounded, company-level knowledge from SEC 10-K filings at inference time, without updating model weights. Within FinLLM-KG, context is supplied either as a natural-language company profile or as a small set of relevance-ranked factual triples. This enables sentiment predictions to account for broader firm-level information rather than the article alone, at almost zero computational cost and for any causal LLM. In addition, attributing each article-level sentiment score to individual sentences and to the retrieved knowledge-graph facts makes the evidence behind the prediction inspectable, thereby enhancing explainability and supporting human-in-the-loop review. Simulations demonstrate that FinLLM-KG consistently improves portfolio performance over the state-of-the-art baseline. ## What Prediction-Market Contracts Reveal at Launch: Trading Volume and Dispute Risk DevFeed: [What Prediction-Market Contracts Reveal at Launch: Trading Volume and Dispute Risk]() Original publisher: [Read original article]() Author: Ahmed Mahrous, Roberto Di Pietro Published: 2026-10-22T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Finance](), [Estimator](), [Data visualization]() Tags: [contracts](), [cs-lg](), [launch](), [q-fin-st](), [risk](), [trading](), [volume]() ### AI overview Using listing-time information for 802,863 resolved Polymarket markets from January 2023 to May 2026, the study predicts lifetime trading volume well in same-period evaluation (out-of-sample R² 0.747 for log volume). Topic contributes most; wording and resolution rules add smaller gains, while neural text models add little beyond interpretable features. Performance weakens sharply across later platform periods: AUC for predicting whether a market reaches $1,000 falls from 0.940 to 0.577, largely amid unfamiliar question formats and changing relationships between contract features and demand. Dispute-risk ranking is more stable (AUC 0.78), but the 0.31% dispute rate yields low average precision (0.012), limiting usefulness for predicting individual disputes; scores are better suited to ranking groups of markets. ### Source excerpt arXiv:2610.29578v1 Announce Type: cross Abstract: Using a near-census of 802,863 resolved markets listed on Polymarket between January 2023 and May 2026, and only information available at listing (the market's title and description, named resolution source, and contract metadata), we predict lifetime trading volume and whether settlement is later disputed. Contract information predicts trading volume well (out-of-sample $R^2 = 0.747$ for log volume), and most of the signal is interpretable: market topic provides the largest gain ($+0.289$ $R^2$ over a $0.220$ contract-feature baseline), wording and resolution-rule design add smaller but measurable improvements ($+0.027$ combined), and neural text models add little beyond the interpretable features. This performance does not survive a period of platform change coinciding with a shift in listing formats: accuracy in predicting whether a market reaches \$1,000 falls from AUC $0.940$ under same-period evaluation to $0.577$ when training on past quarters and testing on later ones. The loss concentrates in markets whose question forms were absent from training (76.9% of test markets; on familiar forms the model still reaches $0.778$), and a decomposition attributes more of it to a changed relationship between contract characteristics and demand than to an outdated mapping from text to features. Dispute predictors are more stable but less precise: launch-time features rank dispute risk far better than chance (AUC $0.78$), yet at a 0.31% base rate precision is too low for case-by-case prediction (average precision $0.012$), so the scores suit ranking groups of markets rather than forecasting individual disputes.