arXiv:2605.16257v1 Announce Type: new
Abstract: Achieving human-level manipulation requires dexterous robotic hands capable of complex object interactions. Advancing such capabilities further demands standardized benchmarks for systematic evaluation. However, existing dexterous benchmarks lack tasks that reflect the unique manipulation capabilities of dexterous hands over parallel grippers, as well as comprehensive evaluation pipelines. In this paper, we present DexJoCo, a benchmark and toolkit for task-oriented dexterous manipulation, comprising 11 functionally grounded tasks that evaluate tool-use, bimanual coordination, long-horizon execution, and reasoning. We develop a low-cost data collection system and collect 1.1K trajectories across these tasks, with support for domain randomization to assess robustness. We benchmark modern models under diverse settings, including visual and dynamics randomization, multi-task training, and action-head adaptation. Through extensive empirical analysis, we identify several important insights and common limitations of current policies in dexterous manipulation, highlighting key challenges for future research in dexterous hand robot learning. Project page available at: https://dexjoco.github.io
Science Journals
arXiv:2602.06239v3 Announce Type: replace
Abstract: We introduce PEPO (Pessimistic Ensemble based Preference Optimization), a single-step Direct Preference Optimization (DPO)-like algorithm to mitigate the well-known over-optimization issue in preference learning without requiring the knowledge of the data-generating distribution or learning an explicit reward model. PEPO achieves pessimism via an ensemble of preference-optimized policies trained on disjoint data subsets and then aggregates them through a worst case construction that favors the agreement across models. In the tabular setting, PEPO achieves sample complexity guarantees depending only on a single-policy concentrability coefficient, thus avoiding the all-policy concentrability which affects the guarantees of algorithms prone to over-optimization, such as DPO. The theoretical findings are corroborated by a convincing practical performance, while retaining the simplicity and the practicality of DPO-style training.
arXiv:2605.16175v1 Announce Type: new
Abstract: Pediatric critical care is a dynamic, high-stakes process involving constant monitoring and adjustments in life-saving treatments. Modeling these interventions is crucial for effective decision support. To address the challenges of high complexity and data scarcity in pediatric Extracorporeal Membrane Oxygenation (ECMO), we frame clinical decision-making as learning to act from trajectories, i.e., imitation learning that learns action models from observational data, with a key feature that actions are not directly observed. We consider TabPFN, a recent transformer-based approach for tabular data, and traditional baselines including XGBoost and Multi-Layer Perceptrons(MLPs) on real-world pediatric ECMO data to learn the action models. We find that the TabPFN-based approach consistently outperforms these classical baselines, supporting its use as a strong clinician-behavior baseline for pediatric ECMO decision support.
arXiv:2605.13925v2 Announce Type: replace
Abstract: Robotic dexterous hands are central to contact-rich manipulation, with rapid progress driven by advances in hardware, sensing, control, simulation, and data generation. However, existing studies are often developed under different assumptions regarding hand embodiments, sensory configurations, task settings, training data, and evaluation protocols, making systematic comparison difficult and obscuring the developmental trajectory of the field. This survey provides a holistic review of dexterous hand research from four complementary aspects. First, we present a hardware-level analysis covering actuation, transmission, perception, and representative hand designs, highlighting the key trade-offs in force capability, compliance, bandwidth, integration, and system complexity. Furthermore, we review control and learning methods for dexterous manipulation from a methodological perspective, grouping representative works by major paradigms and tracing their evolution in chronological order. In addition, we consolidate datasets, modality design, and evaluation practices, which enables methodological progress to be interpreted together with the ways in which it is trained, benchmarked, and assessed. Finally, we discuss the major limitations of current dexterous hand research and summarize the corresponding future directions. By connecting hardware analysis, methodological development, data resources, and evaluation, this survey aims to provide a structured understanding of dexterous hand research and to clarify the most important open challenges for future study.
arXiv:2212.13347v3 Announce Type: replace
Abstract: We report a diffuse Maxwellian illumination scheme for wide-field retinal laser Doppler holography. Inserting an engineered diffuser in the illumination arm transforms a spatially concentrated near-infrared laser focus into an angularly diversified illumination pattern, thereby reducing local irradiance near the anterior segment while preserving coherent interferometric detection. This configuration allows the eyepiece to be positioned closer to the cornea, increasing the digitally reconstructed retinal field of view without producing a localized corneal hot spot. We compare three illumination geometries: focused non-diffuse illumination, diffuse illumination at the same cornea--eyepiece distance, and diffuse Maxwellian illumination. Diffuse Maxwellian illumination expands the retinal field of view while preserving Doppler contrast in broad and high-frequency fluctuation bands. Light-hazard assessment is limited to the current ophthalmic standards ISO 15004-2:2024 and ANSI Z80.36-2021. Based on measured beam profiles, the recommended operating power at 852 nm is set by the most restrictive relevant exposure condition among the assessed anterior-segment, iris, and retinal limits. These results support diffuse illumination as a practical route toward safer, non-mydriatic, wide-field Doppler holography of the human retina.
arXiv:2605.15642v1 Announce Type: cross
Abstract: Nuclear-powered submarines are difficult to track with conventional methods in congested waterways. We revisit antineutrino-based detection as a barrier concept, analogous to a neutrino-enabled SOSUS-style fence in strategic straits. Using analytic scaling relations and numerical estimates, we show that detectability depends primarily on closest approach, detector depth, and deployed mass. For representative assumptions, a 20\,kt detector in the Strait of Gibraltar reaches a local benchmark score $Z_A\simeq2.54$ for an assumed 100\,MW thermal-power sensitivity-study case in a conservative worst-case transit (with Poisson operating point $(P_\mathrm{FA},P_\mathrm{det})\simeq(5.5\times10^{-3},0.51)$ at threshold $k=2$), while a three-detector line raises the mapped score to $Z_A\simeq4.66$. For broad ocean passages such as GIUK, required detector counts are substantially larger; in the baseline maximum passing distance $\mathrm{PDD}_{\max}=5$\,km geometry, about 80 detectors yield only $Z_A\sim1.6$. The paper outlines detector technology choices, statistical assumptions, and deployment constraints for a first-generation feasibility assessment.
arXiv:2603.04299v4 Announce Type: replace
Abstract: Large Language Models (LLMs) often exhibit highly agreeable and reinforcing conversational styles, also known as AI-sycophancy. Although this pattern arises from training objectives that reward user satisfaction over accuracy, it may become problematic when interacting with user prompts that reflect negative social tendencies. Such responses risk amplifying harmful behavior rather than mitigating it. In this study, we examine how LLMs respond to user prompts expressing varying degrees of Dark Triad traits (Machiavellianism, Narcissism, and Psychopathy) using a curated dataset. Our analysis reveals differences across models, whereby all models predominantly exhibit corrective behavior, while showing reinforcing output in certain cases. Model behavior also depends on the severity level and differs in the sentiment of the response. Our findings raise implications for designing safer conversational systems that can detect and respond appropriately when users escalate from benign to harmful requests.
arXiv:2602.06932v4 Announce Type: replace
Abstract: Speculative decoding can significantly accelerate LLM serving, yet most deployments today disentangle speculator training from serving, treating speculator training as a standalone offline modeling problem. We show that this decoupled formulation introduces substantial deployment and adaptation lag: (1) high time-to-serve, since a speculator must be trained offline for a considerable period before deployment; (2) delayed utility feedback, since the true end-to-end decoding speedup is only known after training and cannot be inferred reliably from acceptance rate alone due to model-architecture and system-level overheads; and (3) domain-drift degradation, as the target model is repurposed to new domains and the speculator becomes stale and less effective.
To address these issues, we present Aurora, a unified training-serving system that closes the loop by continuously learning a speculator directly from live inference traces. Aurora reframes online speculator learning as an asynchronous reinforcement-learning problem: accepted tokens provide positive feedback, while rejected speculator proposals provide implicit negative feedback that we exploit to improve sample efficiency. Our design integrates an SGLang-based inference server with an asynchronous training server, enabling hot-swapped speculator updates without service interruption. Crucially, Aurora supports day-0 deployment: a speculator can be served immediately and rapidly adapted to live traffic, improving system performance while providing immediate utility feedback. Across experiments, Aurora achieves a 1.5x day-0 speedup on recently released frontier models (e.g., MiniMax M2.1 229B and Qwen3-Coder-Next 80B). Aurora also adapts effectively to distribution shifts in user traffic, delivering an additional 1.25x speedup over a well-trained but static speculator on widely used models (e.g., Qwen3 and Llama3).
arXiv:2605.12942v2 Announce Type: replace
Abstract: Large-scale datasets have been a key driving force behind the rapid progress of deep learning, but their storage, computational, and energy costs have become increasingly prohibitive. Dataset distillation (DD) mitigates this problem by synthesizing compact yet informative datasets, thereby enabling efficient model training and storage. However, the ease of copying and distributing distilled datasets introduces serious risks of copyright infringement and data leakage. Existing protection methods are primarily designed for raw datasets rather than distilled datasets, and typically rely on backdoor-triggered malicious behaviors, which may raise security concerns. In this paper, we observe that deep neural networks tend to memorize subpopulation distributions during training, resulting in a systematic prediction bias, where models perform better on samples aligned with memorized subpopulations. Motivated by this observation, we propose SubPopMark, a harmless subpopulation-driven protection framework for distilled datasets. SubPopMark consists of two stages. First, the Copyright Verification Marker(CVM) optimization stage injects a class-consistent subpopulation bias while preserving the original optimization trajectory. Second, the User-Specific Tracing Marker (USTM) optimization stage further introduces user-distinguishable perturbations into the CVM-augmented data. To enable black-box verification and tracing, we construct a reference behavior bank by collecting model outputs over carefully designed test sets that cover both standard and subpopulation-shifted data distributions. The provenance of a suspicious model is then inferred by comparing its output behavior signature with the bank and identifying the most consistent reference behavior pattern.
arXiv:2605.12715v2 Announce Type: replace
Abstract: As language models scale, the amount of data they require grows -- yet many target data sources, such as low-resource languages or specialized domains, are inherently limited in size. A common strategy is to mix this scarce but valuable target data with abundant generic data, which presents a fundamental trade-off: too little target data in the mixture underexposes the model to the target domain, while too much target data repeats the same examples excessively, yielding diminishing returns and eventual overfitting. We study this trade-off across more than 2,000 language-model training runs spanning multiple model and target dataset sizes, as well as several data types, including multilingual, domain-specific, and quality-filtered mixtures. Across all settings, we find that repetition is a central driver of target-domain performance, and that mixture training tolerates much higher repetition than single-source training: scarce target corpora can be reused 15-20 times, with the optimal number of repetitions depending on the target data size, compute budget, and model scale. Next, we introduce a repetition-aware mixture scaling law that accounts for the decreasing value of repeated target tokens and the regularizing role of generic data. Optimizing the scaling law provides a principled way to compute effective mixture configurations, yielding practical mixture recommendations for pretraining under data constraints.
arXiv:2605.15920v1 Announce Type: cross
Abstract: We developed a tool for detecting domain shifts, namely subtle differences in the probability distributions of datasets. We identify these shifts using an algorithm designed to detect localised density anomalies in high-dimensional feature spaces. If an anomaly is present, we then identify the feature subspace in which the anomaly is most pronounced. This allows us to trace the domain shift to a small set of features, making the shift interpretable. Moreover, we provide a protocol for compensating domain shifts by extracting, from two unlabelled datasets, subsets of samples with no detectable residual distributional difference. We validate the framework on controlled 20-dimensional benchmarks with known ground truth, recovering both broad and localized shifts together with their supporting feature subspaces. We then apply it to healthy electrocardiogram (ECG) recordings represented by 782 features. In age- and sex-matched cohort comparisons differing in measurement-device composition, the method detects device-induced shifts, extracts representative subsets enriched in the imbalanced device components, and identifies ECG features associated with the acquisition contrast. These results suggest that density-shift detection and subspace attribution provide a practical framework for uncovering hidden cohort biases before downstream modelling.
arXiv:2511.05297v2 Announce Type: replace
Abstract: Digital Adoption Platforms (DAPs) have become essential tools for helping employees navigate complex enterprise software such as CRM, ERP, or HRMS systems. Companies like LemonLearning have shown how digital guidance can reduce training costs and accelerate onboarding. However, building and maintaining these interactive guides still requires extensive manual effort. Leveraging Large Language Models as virtual assistants is an appealing alternative, yet without a structured understanding of the target software, LLMs often hallucinate and produce unreliable answers. Moreover, most production-grade LLMs are black-box APIs, making fine-tuning impractical due to the lack of access to model weights. In this work, we introduce a Graph-based Retrieval-Augmented Generation framework that automatically converts enterprise web applications into state-action knowledge graphs, enabling LLMs to generate grounded and context-aware assistance. The framework was co-developed with the AI enterprise RAKAM, in collaboration with Lemon Learning. We detail the engineering pipeline that extracts and structures software interfaces, the design of the graph-based retrieval process, and the integration of our approach into production DAP workflows. Finally, we discuss scalability, robustness, and deployment lessons learned from industrial use cases.
arXiv:2605.15956v1 Announce Type: new
Abstract: Here we present a massive longitudinal dataset of public Telegram content, comprising over 5.9 billion messages dating from 2015 to 2025, collected from 712 thousand channels and groups, enriched with metadata on forwards, reactions, and polls. The dataset spans multiple languages including Russian and Farsi, representing countries where Telegram shows mainstream adoption, as well as Western languages where Telegram is used in specific sub-communities. The dataset has several advantages. First, when restricted by language, it provides a versatile example of an algorithm-free platform, contrary to many other social media platforms that are strongly influenced by opaque content-curation algorithms. Second, it enables comparative studies across different languages, communities, and user bases under identical platform affordances. The dataset thus offers a foundation for studying engagement patterns, network evolution, and community formation in the absence of algorithmic curation.
arXiv:2605.15656v1 Announce Type: cross
Abstract: Under the trend of multi-waveform coexistence in 6G IoT, intelligent receivers must first identify physical-layer waveform types before performing correct demodulation and resource scheduling. However, existing signal identification research largely focuses on symbol-level modulation classification. Research directly targeting physical-layer waveform types (e.g., OFDM, OTFS, LoRa) is not only extremely scarce but also heavily reliant on deep neural networks and complex time-frequency transforms, making deployment on resource-constrained terminals difficult. Symbol modulation classification methods themselves cannot circumvent the prerequisite of ``waveform identification first.''
To address this dual gap, we propose an ultra-lightweight waveform classification framework based on time-frequency multidimensional features with a cooperative Z-test tree (ZTree). The framework employs low-complexity time-domain feature extraction, and the classification backend adopts a ZTree optimized by Z-statistical testing, which uses hypothesis testing confidence to automatically control decision tree splitting and size, ensuring efficient execution on resource-limited processors. Tested on ten 6G candidate waveforms including OFDM, OTFS, DSSS, LoRa, and NB-IoT, the method achieves 99.5\% average accuracy under AWGN and 87.4\% under TDL-C multipath channels, with main confusion between OTFS and LoRa. Implemented in C on an x86 platform, single inference latency is under 4~ms. To the best of our knowledge, this is the first work achieving real-time recognition of ten IoT waveform types. Future work will target deployment acceleration on embedded MCUs. Code and dataset are open-sourced at: https://github.com/Einstein-sworder/IoT-wave.
arXiv:2602.03922v3 Announce Type: replace
Abstract: Standard sequence mixing layers used in language models struggle to balance efficiency and performance. Self-attention performs well on long context tasks but has expensive quadratic compute and linear memory costs, while linear attention and SSMs use only linear compute and constant memory but struggle with long context processing. In this paper, we develop a sequence mixing layer that aims to find a better compromise between memory-compute costs and long-context processing, which we call online vector-quantized (OVQ) attention. OVQ-attention requires linear compute costs and constant memory, but, unlike linear attention and SSMs, it uses a sparse memory update that allows it to greatly increase the size of its memory state and, consequently, memory capacity. We develop a theoretical basis for OVQ-attention based on Gaussian mixture regression, and we test it on a variety of synthetic long context tasks and on long context language modeling. OVQ-attention shows significant improvements over linear attention baselines and the original VQ-attention, on which OVQ-attention was inspired. It demonstrates competitive, and sometimes identical, performance to strong self-attention baselines up 64k sequence length, despite using a small fraction of the memory of full self-attention.
arXiv:2601.03707v2 Announce Type: replace
Abstract: Existing UAV vision-and-language navigation (VLN) benchmarks rarely provide realistic aerial scenes, natural process-level instructions, and sufficient scale simultaneously, making it difficult to systematically train and evaluate UAV VLN agents under realistic settings. To address this, we propose \textbf{AirNav}, a large-scale benchmark built on real urban aerial data, comprising 137K navigation samples with natural and diverse instructions generated via a human--LLM collaborative pipeline with 10 user personas. We conduct a systematic evaluation of representative approaches on AirNav, ranging from traditional models to multimodal large language models (MLLMs), under unified metrics with open-source implementations. We further propose \textbf{AirVLN-R1}, trained via supervised fine-tuning (SFT) and reinforcement fine-tuning (RFT), achieving state-of-the-art performance with a 51.82\% success rate on the test-unseen split. Real-world experiments on a physical UAV platform provide preliminary evidence of sim-to-real transferability, and our dataset and code are publicly available.
Gene duplication is a major driver of evolution, yet how it generates fundamentally new molecular functions remains poorly understood. Here, we show how such novelty arose in KLMT-1, a selfish toxin that causes genetic incompatibilities in Caenorhabditis tropicalis. KLMT-1 evolved via duplication of an essential tRNA synthetase but, strikingly, lost its ancestral role in tRNA biology and translation. Instead, KLMT-1 localizes to centrosomes, where it targets Aurora kinase A (AIR-1). This innovation is mediated by a three-amino acid insertion that extends a beta-hairpin loop, enabling electrostatic interaction with a regulatory interface on the kinase. Our results demonstrate how changes in selective pressure, combined with minimal modifications in neutrally evolving regions, allow duplicated proteins to access new functional space and evolve entirely new molecular activities.
Holcosus orcesi, the Orces Blue Whiptail, is a Critically Endangered lizard endemic to the upper Jubones River basin in southern Ecuador. Restricted to a narrow elevational range within semi-arid Andean shrublands, it represents one of the few montane members of a predominantly lowland lineage. Here we present the first high-quality reference genome for H. orcesi, generated using Oxford Nanopore Technologies long-read sequencing. The assembly spans 1.68 Gb across only 91 contigs, with an N50 of 76.2 Mb and a BUSCO completeness of 96.8%, making it among the most contiguous and complete squamate genomes to date. Structural annotation predicted 25,682 genes, of which 85% showed homology to known proteins and 45% were assigned Gene Ontology terms. Repetitive elements accounted for 46.3% of the genome, with LINEs representing the predominant class. This genome provides a foundational resource for future evolutionary, comparative and conservation-genomic research of H. orcesi and other mountain reptiles, enabling studies of population genomics, local adaptation, and genomic erosion in isolated populations. By expanding the genomic representation of tropical montane reptiles, this work helps address longstanding phylogenetic and geographic gaps in global biodiversity genomics and provides a foundation for evidence-based conservation of H. orcesi and related taxa.
Gross chromosomal rearrangements are a hallmark of many diseases and cancers. The study of their biogenesis and the mechanisms underlying their formation is greatly facilitated by the availability of genetic reporter assays in model organisms. We present here a novel GCR assay developed in fission yeast, a highly relevant model for understanding genome instability related to human biology. The reporter employs canavanine counter-selection to detect GCRs within a chromosomal context. Using this assay, we identified natural hotspots for GCRs, including inverted long terminal repeats (IR-LTRs). Structural analysis of GCR events showed that IR-LTR-induced GCRs mainly result in either terminal deletions with adjacent inverted duplications or repair via long-range break-induced replication (BIR). Deleting IR-LTRs reduces the GCR rate and reveals another hotspot driven by BIR between homeologous aldo/keto reductase genes on opposite arms of chromosome I. This is the first evidence that BIR can occur in S. pombe on long tracks reaching up to 600 kb. Besides highlighting genome rearrangement hotspots, the assay also identifies regulators of genome instability in fission yeast. Loss of Nup132, a component of the nuclear pore complex, increases IR-LTRs-induced GCRs, while the budding yeast homolog Nup133 has no effect on the stability of a structurally similar IR. In contrast, disrupting djc9, which encodes a conserved histone H3-H4 binding protein, decreases GCR rates. Overall, this sensitive GCR assay enables the identification of factors that control spontaneous and fragile motif-induced chromosomal instability, including those conserved in humans but lost through evolution in other organisms.
Phyllanthus niruri (Phyllanthaceae) is a medicinally important herb known for producing phyllanthin, a bioactive dibenzylbutane lignan with reported hepatoprotective and antioxidant properties. However, the biosynthetic basis of phyllanthin production remains unresolved, largely due to the absence of a reference genome for the species. We report a chromosome-scale assembly of P. niruri generated by integrating PacBio HiFi long reads and Illumina short reads, followed by reference-guided scaffolding against Phyllanthus cochinchinensis. The assembly has an L50 of 7 and 97.6% BUSCO completeness. Annotation predicted 19,254 protein-coding genes, of which 91.1% were functionally annotated, with phenylpropanoid biosynthesis emerging as the most enriched specialized-metabolism pathway in the genome. Using pathway-guided genome mining, structural similarity analysis, and comparative metabolic reconstruction, we propose a putative biosynthetic pathway for phyllanthin originating from the phenylpropanoid-lignan branch through secoisolariciresinol-like intermediates followed by terminal O-methylation reactions. A total of 305 unique candidate genes associated with the proposed pathway were identified, including expanded families of dirigent proteins, peroxidases, secoisolariciresinol dehydrogenases, and O-methyltransferases. Comparative transcriptomic analyses across related Phyllanthus species further supported the proposed pathway through coordinated expression of lignan-associated genes and tissue-specific enrichment of O-methyltransferases. This work provides the first reference genome for P. niruri and a prioritized candidate gene set for functional characterization of phyllanthin biosynthesis.
Cliffs are environmentally extreme yet biodiversity-rich ecosystems that harbour specialist plants, many endemic and threatened. Plant persistence in these nutrient-poor substrates may depend on tightly linked soil- and root-associated microbial communities, which remain poorly understood. These interactions may become increasingly important with the global expansion of recreational climbing. While physical climbing impacts on vegetation are documented, potential chemical effects, from the use of climbing chalk (magnesium carbonate), on soil properties and plant-associated microbiota remain unknown. We sampled soils and roots beneath cliff-specialist and generalist plants, and unvegetated soils, across climbed and unclimbed routes in northern, central, and southern Spain. Soil physicochemical properties were quantified, fungal communities were characterized using ITS-metabarcoding, and structural equation modelling was used to disentangle direct and indirect effects. Climbing increased soil pH and altered soil chemical properties, driving shifts in fungal diversity and functional composition in soil and roots. The relative read abundance of root-associated symbiotrophic fungi declined, whereas arbuscular mycorrhizal fungi and pathogens increased in climbed cliffs. Overall effects were consistent, with cliff-specialist plants mediating nutrient and fungal shifts. Our findings show that climbing can reshape cliff soil chemistry and fungal communities, with potential cascading consequences for plant functional performance, nutrient dynamics, and ecosystem resilience.
Wildlife vaccination could become a powerful strategy to mitigate disease-induced biodiversity losses, yet many vaccines for wildlife diseases provide only limited protection. Notably, tools to control the fungal pathogen Batrachochytrium dendrobatidis (Bd) are urgently needed for amphibian conservation. Laboratory experiments have demonstrated that prophylactic exposure to Bd metabolites increases host resistance, significantly reducing infection intensity in amphibians subsequently challenged with live Bd. Because Bd metabolites are non-infectious and applied topically, this treatment has potential to be administered to waterbodies to vaccinate and protect amphibians. We developed an agent-based model that indicated imperfect vaccination could reduce or amplify Bd infections at the population level, depending on degree of enhanced resistance or tolerance. Utilizing a Before-After-Control-Impact design with ten years of data, we conducted an ecosystem-level trial where we applied low levels of Bd metabolites or a sham control treatment to ponds in California and subsequently quantified Bd prevalence and infection intensity in metamorphosing Pacific chorus frogs (Pseudacris regilla). Unexpectedly, infection intensity was significantly greater in treated ponds relative to control ponds following metabolite addition. Additional model simulations indicated that this could occur via two mechanisms: (1) if treatment greatly increased tolerance alone or in combination with smaller increases in resistance, or (2) if a deleterious environmental interaction caused the treatment to increase susceptibility, rather than promote resistance. Future research is needed to determine whether tolerance or environmental factors drove heightened Bd infection intensities in this field trial to identify contexts in which this treatment can be used as a conservation tool.
Elucidating how habitat degradation facilitates extinction is critical for effective conservation efforts. Here, we propose integrating physiologically-structured population models into stochastic population viability analyses to assess how differing consequences of habitat degradation interact to drive extinction dynamics in a focal population. Using the isolated spectacled caiman Caiman crocodilus population/ecomorph from the Apaporis River as a case study, we find that threatening the resource base, which individuals increasingly rely upon, to outgrow vulnerable size ranges and mature accelerates extinction. We also found that when habitat degradation impacts both the primary adult and juvenile resource bases, this can have marked synergistic effects on threatening population viability. By contrast, destroying nesting sites has only a small effect on accelerating the impact of deteriorating prey availability. Through integrating community-level feedback between habitat degradation/change and population dynamics/structure, our approach provides a comparative framework for assessing the relative importance of distinct mechanisms through which habitat degradation ultimately drives extinction risk.
Although resources are typically distributed continuously in space, species distributions often organize into discrete clusters. In his seminal paper, Turing demonstrated that such clusters can spontaneously arise in population densities, even when populations evolve in environments with continuously varying conditions. This phenomenon is known as Turing instability. In this work, we focus on two models grounded in population dynamics: a one-dimensional model based on the nonlocal Fisher-KPP equation, and a two-dimensional model involving an environmental gradient. We show that phenotypic clusters (sometimes referred to as "species") emerge in these models. We prove that they do not emerge because of Turing instability, but because of stochasticity, and that they disappear when stochasticity is reduced. First, for both models, we start our simulations with initial populations uniformly distributed in the state space. We show that phenotypic clusters quickly emerge and that the distances between them depend on the population size, that is, on the degree of stochasticity. Next, we start from already clearly defined phenotypic clusters. We identify three regimes in the connection between population size, the initial distances between clusters, and the distances between clusters at equilibrium. Last, on the two-dimensional model, we relax the hypothesis of complete clonality by varying the effective recombination rate, explore its effect on phenotypic clustering, and show that phenotypic clustering decays drastically with slight recombination.
Biodiversity is commonly summarized by macroecological mean patterns, most prominently the species-area relationship (SAR) linking habitat area to expected species richness. Yet conservation, policy, and economic decisions increasingly require risk metrics: probabilities of rare but consequential biodiversity shortfalls, including local collapse. Such tail risks are central in finance and insurance but remain difficult to quantify in ecology because the data needed to estimate full richness distributions are rarely available at decision scales. Here we provide a mechanistic route from species-area relationships to biodiversity risk metrics. We show that when regional species abundances are well approximated by Fisher's log-series, a minimal immigration-extinction mechanism yields a closed-form stationary distribution for local richness whose structure tightly couples the mean SAR to richness variability and lower-tail probabilities. This coupling implies exact fluctuation-response identities and an explicit integral transform that reconstructs collapse probabilities and other tail risk measures directly from the mean SAR. These results define ecological analogues of financial risk metrics---such as collapse probability and lower-tail quantiles---without requiring direct estimation of the full richness distribution. Using high-resolution ForestGEO tree censuses spanning tropical, subtropical, and temperate forests, we find empirical support for these predictions across spatial scales. Together, our results show how widely measurable species-area relationships can be elevated from descriptive averages to operational tools for biodiversity risk assessment and reliability-based conservation planning.