Papers

For context-based question answering, using one large model for both evidence extraction and answer reasoning requires substantial memory and computation. Existing approaches trade accuracy for efficiency within a single model only. We propose a multi-model approach for efficient inference: our method, PAIR, employs two differently sized models for evidence extraction and answer reasoning, respectively. A smaller Reader model processes and retains the full context, while a larger Reasoner model sees only the question and generated sequence without directly attending to the context. These capabilities are fused at the logit level: at each token, both Reader and Reasoner produce a prediction, and their logits are combined into one, behaving as a single next-token-prediction model. Fine-tuning further boosts their performance. On HotpotQA, PAIR with a 32B Reasoner retains 96% of that same 32B model's full-context accuracy while using roughly one quarter of its KV memory. At 45.7K-token contexts, PAIR achieves up to 3.2$\times$ higher end-to-end throughput than the full-context 32B model on the same GPU. PAIR also improves the accuracy-memory and accuracy-throughput trade-offs over evaluated token-reduction baselines. Its accuracy-preserving behavior generalizes across Qwen2.5, OLMo-3, and Gemma-3 model families.

DNA methylation is a critical epigenetic modification that regulates gene expression and genome stability, primarily catalyzed by the DNA methyltransferase 3 (DNMT3) family of enzymes. Distinguishing precise sequence-specific flanking preferences between DNMT3A and DNMT3B remains computationally challenging due to the complex and high-dimensional nature of biochemical sequence data. In this study, we propose a novel Quantum Machine Learning (QML) approach to classify DNA methylation preferences, leveraging a Variational Quantum Classifier (VQC). Using experimentally derived deep enzymology data, we constructed a dataset containing 4,096 unique NNNCGNNN motifs. We evaluated a four-qubit VQC using a custom NumPy exact-statevector simulator. The described two-layer circuit contains 24 trainable rotations, eight CNOT gates, and eight input-encoding rotations, totaling 40 gates before measurement. Across five initialization seeds, the revised VQC evaluation achieved a ROC-AUC of 0.661 pm 0.009, where variability denotes sample standard deviation. The held-out test partition is approximately balanced, containing 200 DNMT3A-labeled motifs and 210 DNMT3B-labeled motifs. An always-DNMT3B predictor achieves 51.2% accuracy and a balanced accuracy of 0.500 on this test partition. This proof-of-concept establishes a foundation for utilizing near-term quantum architectures in complex epigenetic classification tasks, paving the way for hybrid quantum-classical pipelines in bioinformatics.

Software-defined radios (SDRs) have become an inexpensive, portable platform for replay attacks, location spoofing and jamming. The HackRF One can operate stand-alone when fitted with a PortaPack interface board running the open-source Mayhem firmware. No peer-reviewed study has addressed the forensic analysis of this device family. This work establishes, at source-code level, where forensically relevant data reside on the microSD card of a device running Mayhem v2.0.1, in which format, when they are written and how they should be interpreted. The methodology is applied to an SD card image that the author acquired from their own device after four months of real use, and every inference is drawn from the image alone. The workflow covers timestamp rules that depend on the real-time clock (RTC) battery state, traces of on-device trimming and text editing, cluster-level recovery of deleted in-phase/quadrature (IQ) captures, and decoding of recorded signals cross-checked with an independent decoder. In the example image, the final part of a deleted IQ capture and a trimmed-away capture segment are recovered, and six captures are attributed to a single rolling-code on-off keying (OOK) transmitter sharing one 32-bit identifier. The study provides practitioners with an artifact catalogue, interpretation rules and acquisition recommendations.

In this paper we provide an introduction to some of the efficient features of the Raspberry Pi Pico (RP2) family of microconroller boards (RP2040 and RP2350). More exactly, we show how to use PIO (Programmable Input/Output) and DMA (Direct Memory Access) to program the RP2 boards for efficient DAC (Digital to Analog) and ADC (Analog to Digital) conversions. Using OP-AMPS (Operational Amplifiers) we also show how to build a minimal electronic frontend interface for signal input/output. Finally, we integrate all these components into a product, the O-SCOPE, which is a practical and easy to use oscilloscope and wave function generator. The RP2 component of the O-SCOPE is built using Micropython, while the user interface is developed using vanilla Javascript and it runs on the web browser of any computer, interacting with the RP2 board via the Web Serial API. The O-SCOPE web interface is hosted on GitHub Pages such that it is directly accessible everywhere using a web browser, such that no local installation is required. Also, the RP2 software component is freely distributed on GitHub using the UF2 (USB Flashing Format) mass storage file, and it can be installed by simply dragging the file on the RP2 boards. These features make the O-SCOPE easily accessible, transforming it into a very low cost and useful tool for laboratories everywhere in the world.

Entanglement entropies of multipartite pure states satisfy linear inequalities; for stabilizer states, the backbone of quantum error correction, the achievable entropy vectors generate a convex cone. We show that, after a linear lift, the closure of this cone is exactly the self-dual slice of a classical object on one more element than there are parties: the closed cone generated by the rank functions of arrangements of subspaces over $\mathrm{GF}(d)$ for a prime local dimension $d$, and of subgroups of abelian groups of exponent dividing $d$ in general. This correspondence brings two features of rank inequalities into play. First, the purifier acts as an extra variable: already for five parties the stabilizer cone is strictly smaller than the quantum linear rank cone, with which earlier evidence suggested it might coincide; an explicit six-variable inequality witnesses the gap. Second, the characteristic of the field enters, so the cone depends on the local dimension. It does not for up to four parties. With enough parties, the cone of $\mathbb{Z}_d$ qudits is contained in that of $\mathbb{Z}_{d'}$ qudits if and only if $d$ divides $d'$, and already for six parties the logical zero state of Steane's code lies outside the stabilizer cone of every odd local dimension. For this last result we show that the seven-variable characteristic-dependent inequalities we checked all follow from Shannon inequalities on the self-dual slice, and instead transfer one of them through a family of subgroups derived from the Fano plane.

A reading departs from what a monitoring system expects for one of three reasons: the physical process changed, the instrument is failing, or the value was corrupted between the instrument and the software. The three call for different responses and look alike in a single sample. We state the three-cause problem, prove which pairs of causes one channel cannot separate, and propose a bank of online detectors - a Kalman tracker with CUSUM and EWMA charts on its innovation, a Hampel median with generalized-ESD confirmation, a robust noise statistic with ADWIN, exact-value checks and an optional two-channel complementary filter - combined by a deterministic, inspectable attribution rule. The detectors were merged into a public algorithm library. On 2,942 injected events of thirteen fault types the method detects 96.3% (95% interval 95.7-97.0) at 0.22 false alarms per hour and attributes 91.2% (90.6-91.8) of detected events correctly from one channel and 99.0% (98.6-99.4) with a second, rate channel; from one channel only 2 of 207 detected sensor drifts are attributed to the instrument, as the problem statement predicts. Injected into recorded industrial signals, the same catalogue is detected at 80.8% and attributed correctly at 84.5%, at 3.6 false alarms per hour. On the Skoltech Anomaly Benchmark the method detects 32 of 33 experiments with a median delay of 15 s, at 46 false alarms per hour against 11 for a rolling z-score.

Thompson sampling (TS) achieves logarithmic and asymptotically optimal regret in classical stochastic bandits, but its behavior in multi-agent learning remains less understood. We study TS in an unknown finite two-player zero-sum matrix game with Gaussian bandit feedback. Under centralized self-play, TS samples a single payoff matrix from the posterior and plays a Nash equilibrium of the sampled game. For every fixed zero-sum game with a unique Nash equilibrium, we show that the expected cumulative duality gap is $O(\sqrt T)$. The rate improves to $O(\log T)$ for pure equilibria. We then identify the exact coefficient of the $\sqrt T$ term, which depends only on the equilibrium strategies and the noise level. We also establish a minimax lower bound matching the square-root dependence on $T$ and an instance-dependent lower bound matching the logarithmic dependence on $T$ for a class of games with a unique pure equilibrium. Experiments support the predicted rates and coefficient.

Knowledge distillation transfers representational knowledge from a large teacher network to a compact student. Existing relational and contrastive distillation methods treat all samples uniformly, spending equal gradient on well-matched and poorly-matched regions of the representation space alike. We propose Representational Difference Distillation (RDD), which diagnoses where teacher and student representations structurally disagree and concentrates the distillation signal on those regions. RDD uses two asymmetric affinity matrices that capture opposing directions of misalignment: one identifies samples the student incorrectly considers similar, the other identifies teacher relationships the student has yet to learn. These affinities determine which sample pairs the distillation loss aligns and how strongly it corrects each misaligned sample. On CIFAR-100 with resnet32x4 to resnet8x4, RDD reaches 75.58% test accuracy in our runs, on par with CRD (75.45%) and above KD and RKD. We additionally explore curriculum-based and sampling-based variants, finding that weighting the loss is more effective than filtering the data.

In this paper, we propose the Dissipative Lagrangian Neural Network (dLNN), a continuous-time system identification technique that separately learns the conservative and dissipative dynamics of a mechanical system directly from data. A key contribution is a novel parametrisation of dissipative forces that (i) is structurally constrained to be dissipative, enforcing passivity of the learned model, (ii) cannot absorb conservative mechanics, and (iii) retains sufficient expressivity to represent a broad family of state-dependent damping and regularised friction forces. The model is validated on prediction tasks on two real-world mechanical systems: trajectory prediction of a chaotic double pendulum and inverse-dynamics modelling of a six-degree-of-freedom industrial robot. Across these tasks, the dLNN demonstrates improved predictive performance over existing Lagrangian architectures while remaining competitive with a black-box neural model. The imposed physical structure also improves generalisation to unseen states, with the benefits becoming particularly pronounced when training data is limited. Finally, we introduce a symbolic distillation procedure that extracts closed-form expressions from the learned physical components. Applied to the industrial robot, this produces interpretable symbolic models that retain high predictive accuracy while reducing the computational overhead associated with evaluating neural Lagrangian models.

This paper presents AutoSec-RAN, a specification-grounded large language model (LLM) framework for automated security triage of Open Radio Access Network (O-RAN) testing. Unlike approaches relying on static rule sets or privileged access to raw packet captures, AutoSec-RAN fine-tunes Mistral-7B-Instruct through two Low-Rank Adaptation (LoRA) stages over a curated corpus of 69 O-RAN, 3GPP, and adjacent standards documents, then selectively augments inference with retrieval-augmented generation (RAG) over the same specification index. At inference, the model receives only an investigator-style question and a scenario metadata block, simulating an automated analyst operating without raw packet evidence. AutoSec-RAN is validated on a purpose-built 5G/O-RAN testbed across labeled fronthaul distributed denial of service replay scenarios. A Gemma-4-31B judge LLM scores each candidate's answer on faithfulness, correctness, and actionability relative to a reference authored from full packet capture and log evidence. AutoSec-RAN achieves a mean judge overall score of 0.713 across all 313 evaluation items (0.755 restricted to the five supervised taxonomy classes), clearing the 0.70 reporting threshold, and a blended score of 0.778 (0.808 supervised-only) when cosine similarity over Beijing Academy of Artificial Intelligence (BAAI) General Embedding (BGE) is incorporated as a secondary semantic signal. The framework performs strongest on mitigation and threat-model reasoning (with a mean overall of 0.849 and 0.848, respectively) and identifies a measurable gap on the held-out attack-direction class (0.515), a controlled result that exposes a specific training data requirement. These results establish specification-grounded LoRA fine-tuning combined with RAG retrieval as a viable foundation for automated security validation of AI-native O-RAN.

Representing individual-level behavioral heterogeneity is challenging in epidemiological modeling, particularly when empirical behavioral data are sparse. We assess the potential of large language model (LLM)-generated behavioral outputs in addressing this challenge, using hepatitis C virus-related behaviors among people who inject drugs as a case study. Correspondence with epidemiological patterns was evaluated using external benchmarks: literature-informed association directions, correlations from an agent-based model (ABM), and survey associations. Across three LLMs, unconditional directional recovery across 28 retained literature-informed associations was 46-68%, compared with 82-93% agreement when LLMs were directly asked to classify the same associations. Mean absolute differences from ABM and survey correlation patterns were 0.11-0.13 and 0.15-0.18, respectively. Together, these findings show partial alignment between LLM-generated behavioral relationships and epidemiological evidence, despite higher knowledge-probe agreement. More systematic evaluation is needed before incorporating LLM-generated behavioral outputs into epidemiological models that simulate transmission dynamics.

We study how initial model selection, later handoff, and training targets affect success and cost in LLM tasks that span multiple turns. We compare several router architectures using handoff curves, which record task success and cost as the handoff step changes. Our dataset contains 34,204 complete executions of three Qwen models across BFCL, MultiWOZ, and $\tau$-retail, providing offline supervision for routing decisions. In our primary comparison, with Qwen3.5 35B-A3B as the starting model, allowing a later handoff does not establish an improvement in utility over initial selection on BFCL or MultiWOZ, and increases cost on BFCL across a range of cost weights. Changing the supervision targets improves router performance on both our dataset and TwinRouterBench. On our dataset, replacing noisy training success curves with monotone fits raises success by 0.2 to 4.1 percentage points at the same cost across the reported operating points. On TwinRouterBench, we construct targets that describe the model capability needed for the remaining work and difficult steps ahead. Compared with training on the benchmark's per-step labels, these targets raise trajectory pass scores by 7.8 percentage points and cost savings by 6.3 percentage points on the full static bank. A live deployment of the router trained with these targets resolves 79 of 100 instances at a router cost of 0.59 US dollars per instance.

Distributed agent workflows cache delegated authority at tool gateways. Revocation may be reported at issuer commit, transport acknowledgment, or receiver persistence, although none of these observations establishes that all gateways enforced the change or that guarded effects stopped. We define C0-C5, a bounded outcome contract spanning issuer commit through effect prevention, and implement it as a persistent monitor over a frozen gateway set, call set, and observation window. A 14-case fault matrix checks conservative classification, sticky violations, restart persistence, and evidence immutability. In a derived audit of its eight semantic cases, treating C3 or C4 as completion produces seven or five false positives, respectively; the bounded C0-C5 verdict produces none. Descriptive scaling covers 1,024 gateways and 10,000 effect vectors. The primary experiment uses the maintained Keycloak 26.7 identity server and measures the contract across two durable receivers, eight physical agent-tool gateways, Filesystem and PostgreSQL effects, and 16 AWS hosts in three availability zones. Normal delivery and recovery commit no effects. Under receiver-selective non-delivery, transport acknowledgment and receiver persistence still succeed, but 112 of 225 calls (49.78%) are admitted, and all 112 effects commit in every treatment block. A separately provisioned execution repeats the result. OPA/Toxiproxy and a live three-role LLM workflow with networked MCP tools reproduce the persisted-but-not-enforced boundary on narrower surfaces. Finally, an unmodified external MCPMark agent executes 60 closed-loop task arms. Normal C4 denies all 31 exposed post-barrier calls and commits no effect; selective C3 allows 344 calls and commits 145 effects across Filesystem and PostgreSQL. Transport success is therefore evidence of progress, not a revocation-completion verdict.

Authorization checks at individual tool calls do not reveal how long earlier authority remains usable. We study this lifecycle failure as privilege creep: a capability stays reachable after its last justified workflow step. We present CreepScope, a trace-oriented evaluation protocol that samples reachable authority at every step and separates no-remaining-use exposure into stale-after-use and never-required components; Capability-Step Exposure (CSE) counts total stepwise overexposure and Residual Privilege Debt (RPD) the terminal residue, while stale-after-use exposure is the direct privilege-creep measure. A 96-cell deterministic replay over 12 controlled storage, communication, and cloud-operation workflows validates these semantics: Static Task Scope and Monotonic On-Demand retain mean CSEs of 26 and 13, whereas an oracle-controlled Step Lease has zero CSE and zero residual debt by construction. All three persistent regimes share the same stale-after-use exposure of 13, so their CSE differences come from scope and grant timing rather than post-use retention; Full Access alone adds 91.33 never-required capability-steps. A separate 24-episode local case then lets Qwen2 7B choose tools in one simulated mail workflow: the injected stale send executes in all nine persistent-policy episodes and in none of the three Step-Lease episodes, while all 24 required action sequences complete. These results establish a reproducible containment case, not cross-workflow or cross-model effectiveness. Core metrics are reconstructed from raw events, and exact gates reject incomplete or duplicated evaluation grids.

Muscle fatigue limits the effectiveness of functional electrical stimulation (FES), yet fatigue monitoring during active stimulation remains challenging because electromyography (EMG) is severely contaminated by stimulation artifacts. To address this issue, this study proposes an integrated mechanomyography-based sensing framework (MMG-FES) for artifact-resistant fatigue monitoring during stimulation. Fatigue states were first characterized during voluntary movement using established EMG spectral indicators, including median frequency (MF) and mean power frequency (MPF), and then linked to MMG-derived features, including high-frequency energy ratio (HFER) and spectral bandwidth (BW), for fatigue recognition under stimulation conditions. Experimental results showed consistent fatigue-related trends, with EMG spectral features decreasing and MMG features exhibiting coordinated spectral redistribution across repeated contractions. Using these features, the proposed framework achieved fatigue-state recognition accuracies of 0.93 and 0.96 with logistic regression and random forest, respectively. These findings support MMG as a physiologically meaningful and stimulation-compatible modality for fatigue-aware closed-loop neuromuscular stimulation.

In this work we present a probabilistic model for predicting the temporal evolution of wildfire size using conditional generative algorithms, namely flow matching. Rather than predicting spatially resolved fire progression, the proposed model forecasts a time series of fire acreage, substantially reducing computational cost and enabling efficient integration with downstream applications such as early warning systems and risk assessment tools. Given estimates of the fire size at the current and previous forecasting steps, alongside spatially averaged weather conditions, terrain characteristics, and fuel composition, the model predicts the change in fire area over a 3-hour interval. Training data are drawn from 140 coupled atmosphere-wildfire simulations generated using WRF-SFIRE, paired with weather data from the North American Mesoscale (NAM) model. Sensitivity analysis demonstrates that the model captures physically-consistent responses to variations in wind speed, temperature, and relative humidity. Model performance is evaluated against reference solutions from WRF-SFIRE. For single-step predictions, the model achieves a symmetric mean absolute percentage error (sMAPE) of 0.39 relative to the reference WRF-SFIRE solutions. Recursive multi-step predictions over 24 hours are performed for 12 held-out test cases, where the flow matching model achieves a final fire area sMAPE of 0.39 while further providing meaningful quantification of forecast uncertainty.

Emotion datasets are routinely mapped onto positive and negative valence, by assigning each emotion category one polarity. We test where this works and where it fails, using SemEval-2018 Affect in Tweets (English, Arabic, Spanish), Affective Text aligned with EmoBank, raw GoEmotions annotations, naturalistic hashtag corpora, and a controlled annotation study. Fixed projection works well when a text carries labels of one valence. It fails on a subset identifiable from the labels alone: tweets whose emotion labels point in opposite directions. These are 9.0% of polar SemEval test tweets but cause 54.8% of the polarity errors of a learned projection, are more than twice as likely to receive a neutral-or-mixed gold label, and are not rescued by modeling the full label set. Annotators produce such configurations routinely: 16.3% of multilabel GoEmotions judgments combine emotions from the dataset's own positive and negative parents. In a controlled study holding word and reader constant, context changes the coarse valence judgment in 56.8% of cases, against 19.9% for controls. We recommend projecting label configurations rather than labels, abstaining or keeping a mixed class where labels conflict, and reporting coverage alongside accuracy.

This paper investigates a new role of randomness in stochastic policies–as a medium for embedding and communicating information. Specifically, we consider an agent interacting with an environment according to a prescribed stochastic policy. By carefully modulating the random action realizations of the policy, information can be embedded into the agent's behavior without altering the prescribed action distribution. We cast this problem as a channel coding problem, in which the controlled system implicitly serves as a communication channel. We then propose a coding scheme, termed posterior-matching randomness modulation (PRAM), that selects actions according to the message to be transmitted. With the aid of common randomness independent of the control system, PRAM preserves the prescribed action distribution for every state and every message. We further prove that PRAM is asymptotically reliable: the message can be decoded with arbitrarily small error probability as the observed state trajectory becomes sufficiently long. Experiments demonstrate the effectiveness of the proposed framework in applications including multi-bit policy watermarking and trajectory provenance.

Full-support local randomizers can incur unnecessary reporting distortion on ordered alphabets. We study randomized response whose output is restricted to a radius-$t$ neighborhood of the private input. Input-dependent supports create privacy leakage at their boundaries, so ordinary likelihood-ratio calibration is no longer sufficient. Under approximate local privacy restricted to input pairs at distance at most $H$, we derive the exact hockey-stick defect of the mechanism. We then solve the joint design of the truthful probability, off-truth probability, and support radius. The fixed-radius optimum is available in closed form, while the global optimum requires only a one-dimensional integer search. We characterize both the privacy floor and its inverse, namely the smallest defect attainable under a reporting-distortion budget. We further prove that stringent approximate privacy forces the support size to scale as $\Omega(H/\delta)$. Deterministic evaluations illustrate the resulting tradeoff.

In multi-agent reinforcement learning, exposing a cooperator to partners with diverse strategies during training is key to better collaboration at test time. However, existing methods typically rely on trajectory-level proxies to measure diversity, which may label two distinct trajectories of the same strategy as diverse. Instead, our key insight is to leverage LLMs' ability to describe behaviorally distinct agents in environments, similar to how humans can recognize high-level strategies. To this end, we introduce PopGen, a framework that synthesizes semantically diverse agents from scratch using LLMs. PopGen first generates a diverse set of textual descriptions of agent behaviors, then translates each behavior into a code policy that can be executed within the environment. We show that cooperators trained against PopGen populations achieve robust zero-shot cooperation with unseen partners, outperforming population-play baselines across four collaborative environments. We further show that semantic diversity scales with PopGen population size, improving the cooperator's worst-case performance, and we identify design choices that preserve population quality while maximizing diversity. Our code is available at https://github.com/labicon/PopGen.

Spiking Transformers achieve energy efficiency by discarding softmax normalization in favor of thresholded, binary attention. However, this removal breaks the thermodynamic framework used to analyze attention networks, as the normalized measure and partition function cease to exist. We restore this theoretical foundation by proving that softmax-free spiking attention operates exactly as the zero-temperature limit of a product of independent two-level systems. This correspondence places spiking attention strictly in the frozen ordered phase of the thermodynamic phase diagram for all inputs. Consequently, the entropic cost of organizing attention collapses, replaced by a Landauer-type physical floor governed by action-potential generation. Furthermore, this ordered-phase pinning mathematically explains the optimization difficulty caused by vanishing surrogate gradients. We validate this theory with forward-relaxation experiments, placing both standard and neuromorphic attention mechanisms on a unified thermodynamic axis.

We present a low-cost miniature Ackermann vehicle for embedded neural-driving research, organized as a hardware-partitioned perceive-decide-act agent: an ESP32 camera node streams frames over Wi-Fi to a host computer, where neural inference and state estimation run, while an ESP32-S3 chassis controller executes safety-critical actuation in a 100 Hz per-wheel proportional-integral-derivative (PID) loop beneath a micro-ROS client. The partition isolates real-time control and safety on the vehicle while learning-based components run off-board, supervised by an operator from a gamepad. We characterize the complete pipeline during real closed-loop neural driving. Wireless sensing, rather than neural inference, dominates latency, with a 59.5 ms median capture-to-actuation delay; the platform absorbed wireless variability by holding the last safe command under a firmware watchdog, operating across twelve autonomous sessions with zero collisions or manual interventions; and encoder-inertial-measurement-unit (IMU) fusion cut odometry loop-closure error by 22%. These results indicate that carefully partitioned microcontroller-class vehicles support reproducible experiments in embedded AI, wireless control, and operator-supervised autonomous driving. The platform software is open source.

We study pretraining when compute is abundant but unique training data are finite, requiring repeated training on the same data. We first revisit infinite-compute pretraining (Kim et al. 2026) and show that its frontier depends critically on choices previously fixed or studied only in isolation. In particular, optimizing batch size together with epoch counts reveals a U-shaped relationship between validation loss and batch size, favoring smaller batches than typical pretraining, likely because small-batch gradient noise provides implicit regularization. We further find that training recipes, previously optimized under a single-pass assumption, can be tuned, e.g., repacking documents into new sequences every epoch, weight tying, and positive embedding weight decay. We then relax the infinite-compute assumption. In particular, smaller-than-typical batch sizes can require prohibitively long sequential training, motivating to treat GPUs and time as separate resources. Tighter deadlines favor larger batches, and latency-oriented parallelism can trade additional GPUs for shorter training. Nevertheless, even under a fixed time budget, a smaller batch with fewer GPUs can achieve lower validation loss than a larger batch with more GPUs, challenging the common practice of increasing data parallelism to maximize GPU utilization under a deadline. Together, our results identify underexplored factors in repeated-data training and provide concrete guidance on how training recipes and resource allocation should change from single-pass pretraining.

A self-evolving agent rewrites its own harness and must decide which rewrites to keep, so its improvement signal is the check a rewrite has to pass. Verifying a rewrite by agent rollout is expensive and subjective, while executable checks such as the repository's own tests are too sparse to catch behavior-changing rewrites, rejecting only a quarter of them. We instead take test feedback as the executable signal, and search for test suites that are both small and strong with a quality-diversity evolutionary algorithm, so that the signal is dense and informative. TeSearch treats suite size as a behavior descriptor, so that quality is pursued at every size and a small, strong suite is reachable, and learns a transition kernel during the search that denoises the LLM as an actuator on that descriptor. Across 129 real-world modules and three backbones, TeSearch delivers the highest-quality suites of any method compared, LLM-driven evolutionary search included, and the smallest of any LLM-based one. Used as the improvement signal of a self-evolving agent, it rejects 59.9% of behavior-changing rewrites against 25.3% for the repository's tests, at a 4.0% false-alarm rate and 3.5 s per decision. The code and data are available at https://github.com/ccsnow127/TeSearch.

High-pileup conditions in CMS during the HL-LHC era make charged-particle tracking increasingly challenging as detector occupancy and combinatorics grow. We present a hybrid approach that exploits Line Segment Tracking (LST) objects rather than individual hits to enable the first CMS ML-based track reconstruction algorithm. The LST objects, namely the triplets used in this study, are built according to geometry- and physics-driven criteria and carry richer local structure, geometrical cues, and physics quantities such as transverse momentum, which helps the model incorporate detector-level information from the start. We apply a Transformer encoder with geometrically masked sparse self-attention to embed these objects in a learned latent space, and use object condensation to assemble full tracks in a single inference step. This object-based representation reduces complexity in dense environments and makes the model less sensitive to ambiguities introduced by high pileup. Evaluated on CMS Phase-2 simulation of $t\bar{t}$ events at $\langle \mu \rangle = 200$, the approach achieves reconstruction efficiency comparable to the LST quintuplet baseline while reducing the fake rate by roughly a factor of three and suppressing duplicates almost entirely, demonstrating the promise of advanced ML architectures for next-generation particle tracking.

Emergent misalignment (EM) occurs when fine-tuning a language model to a narrow task makes it misaligned on unrelated ones. Because organizations now routinely fine-tune models on their own data, it matters which kinds of training data cause EM. So far, studies have induced EM with data containing recognized safety failures, such as insecure code or harmful advice, while biased training data is typically examined by fairness research for the bias models learn from it rather than for its effects on safety. In real-world text, much of this bias, such as demographic stereotypes, appears not as derogatory language but politely, embedded in judgments about people. We ask whether learning such polite stereotypes can itself cause EM. We find that fine-tuning on polite answers that rely on stereotypes makes models less safe and less truthful on unrelated tasks. The effect holds across four open-weight model families (7B to 12B), both gender and race stereotypes, and repeated training runs, relative to controls trained on the same questions without the stereotype. We further show that the learned stereotype has its own direction in model activations, distinct from directions for generic misaligned traits such as deception. This direction exists before fine-tuning and can both control the learned stereotype and monitor its acquisition. Existing defenses against EM improve safety while leaving the stereotype largely intact. Beyond this controlled setting, fine-tuning on political news coverage also triggers EM, apparently amplified by partisan framing, and stereotype models acting as tool-using agents can carry out harmful tasks and blame people without evidence. Our results show that biased training data is not only a fairness problem but also a safety one. We argue that EM is studied too narrowly, and that fairness concerns and methods offer a natural way to broaden the search for its triggers.

Text-conditioned symbolic music generation has attracted growing interest with the development of generative models. Yet two questions remain underexplored: what aspects of music the conditioning text should describe, and how the corresponding data should be collected and trained to improve controllability and text-music alignment. We address these challenges with TASTE, a taxonomy-aware framework for symbolic music generation with a main focus on piano pieces. First, we introduce a taxonomy of musical elements that specifies how a piece can be captured through textual description in a musically grounded way. The taxonomy comprises a set of metadata tags (genre, emotion, section plan, etc.) together with a free-form description of music imagery. This structure decomposes textual input to match the generation goal more coherently and precisely. Second, we build a data pipeline for collecting, annotating, and retrieving public symbolic piano resources, yielding a corpus of 432K MIDI files labeled at multiple levels of granularity. Finally, we design a staged training procedure that leverages this multi-granularity data across two model families: an autoregressive model and a non-autoregressive diffusion model. Both objective and subjective evaluations show that our models achieve superior musical-element control, text-description alignment, and overall music quality. We will release the full pipeline, all annotated datasets, and the model checkpoints.

Transformer models trained on symbolic music have become remarkably capable. However, it is yet unclear how they represent music structure, and whether understanding their internal representations can be used to unlock methods for model control. We examine this through a case study of harmonic analysis: when a music transformer navigates a cadence, what does it internally "know" about its harmonic context, and can that knowledge be used to affect the generation? Using Aria, a LLaMA-based symbolic music transformer evaluated on Bach chorales, we find that harmonic decisions are causally localized progressively from middle layers onward. Linear probes show that mode becomes decodable in the same layer range where patching identifies growing causal effect, while key identity emerges later. We then confirm these representations are actionable for music generation control: steering via latent directions corresponding to harmonic changes reliably produces a deceptive or minor resolution, while preserving structural coherence and generation quality. This demonstrates that, at inference time, the model's internal harmonic knowledge can be directly manipulated to control the harmonic output.

Vision-language-action models for embodied navigation typically rely on RGB observations, limiting their applicability in lights-out industrial environments where visible illumination may be unavailable. We study illumination-independent navigation using spatial audio for target localization and depth for obstacle-aware geometry. To support this setting, we introduce DarkNav-Bench, a benchmark with 24 industrial scenes and four progressively challenging acoustic conditions. We further propose ADAPT, an audio-depth action-planning framework equipped with spatial-audio perception and optimized for closed-loop navigation through reinforcement learning (RL). Our key RL contribution, transition-aware reward shaping (TARS), defines a pose-aware shortest-action potential over a position-heading graph. Unlike conventional geodesic-progress rewards, TARS accounts for both feasible movements and in-place rotations, providing action-consistent feedback for policy optimization. Experiments show that ADAPT consistently outperforms training-based navigation baselines and zero-shot multimodal models on both seen and unseen scenes. On unseen environments, it achieves 90.6% success under Clean conditions and retains 60.0% success under the most challenging Localized Noise setting. These results demonstrate the potential of spatial audio and depth as complementary modalities for reliable RGB-free navigation in lights-out industrial environments.

A multiple-access channel transmits messages from several senders to one receiver. For quantum outputs, the receiver must distinguish states depending on all messages, while tests for different decoding errors may not commute. In this paper, we give a self-contained account of simultaneous decoding, in which one joint measurement recovers the messages at the established achievable rates using independent codebooks. The proof combines Sen's tilted-subspace method with Cheng's analysis of the pretty-good measurement, a standard measurement for distinguishing quantum states. Auxiliary labels close to a common reference state separate the geometric directions associated with different errors. A Gram-matrix bound then allows separately chosen binary tests to certify the performance of one measurement. The labels are removed from the completed decoder, leading to a deterministic packing theorem that applies to finite state ensembles, including unequal priors. We develop the two-sender case first, with worked examples and figures, before treating any fixed number of senders. Sharper finite-error guarantees for the same decoder are developed separately, with sharp bounds for two specified steps of the analysis rather than a claim of globally optimal decoding. Applications recover known achievable rates for unassisted and entanglement-assisted classical communication, distributed classical compression with quantum side information, and coherent-state communication through a beamsplitter or passive multiport device. The common proof makes the geometric mechanism and its quantitative consequences available without repeating the construction for each application.