Matthew Edward Scarborough, Heinz Furrer, Torsten M. Scheyer
et al.
Here we report on the anatomical re-investigation of a Late Pleistocene (Middle Würmian Interstadial, MIS 3, ca. 40–50 ka BP) perinatal woolly mammoth excavated in 1890-91 from Niederweningen, Switzerland, the first perinatal mammoth ever scientifically studied. In order to narrow down its age at death we compare perinatal woolly mammoths and African elephants in terms of their (i) ontogenetic allometry, especially in woolly mammoths from Russia, (ii) bone microstructure, (iii) tooth eruption sequences, and (iv) the timing of cranial suture closure. A distinct layer of postnatal bone deposition in the tibial midshaft and maxilla identified using micro-CT scanning, in particular, provides strong evidence of a post-natal ontogeny despite the lack of obvious wear in the dentition, suggesting an age of no more than a few months. Odontochronologically, the individual probably belongs to Jachmann’s (1988) Dental Stage 1 (3 months - 2 years), and tiny pits in the enamel/dentin tips of lamellar edges are analogous to those seen in juvenile M. primigenius from Russia (“Lyuba”), and are variously hypothesized to be caused by initial wear, spalling, dissolution pitting, incomplete mineralization, or eruption through the gums. Additionally, we describe anisomelia of uncertain aetiology, resulting in the stunting right fore-leg, perhaps contributing to the cause of its death. Finally, we discuss the significance to ecology and extinction of an altricial life-history strategy and the relatively rapid post-natal growth evidenced in woolly mammoths, a taxon with an exceptionally long gestation period.
A defining feature of the hominin clade is bipedality, often parcelled together with terrestriality. However, there is increasing evidence of locomotor diversity, both within the hominin clade and amongst the Miocene apes that came before them. There is also growing recognition that bipedalism might have arboreal origins and that arboreality persisted in several hominin taxa, including our own genus Homo. Furthermore, the difference between terms like “habitual” and “obligate” bipedality is not clearly defined and is often inferred from fossil features, rather than a description of each behaviour in vivo. Combining fossil and palaeoecological evidence with insights from behavioural ecology facilitates new interpretations of evolutionary pathways and highlights the importance of considering convergent evolution in the emergence of locomotor traits and characteristics. Taking such an approach also moves away from assumptions of a straight-line trajectory towards modern human locomotion and explores the likelihood that independent forms of bipedality and terrestriality arose at different times and in different combinations with other features of ape morphology and behaviour. Evidence from extant primate species can broaden our understanding of the correlates, causes, and consequences of terrestriality and can be used to generate hypotheses which are then explored further using paleontological methods. In this paper, we explore the evolutionary origins of hominin locomotion, but extend our review to include broader timescales, a wider range of primate taxa, and an integrated set of methods and disciplines for generating and testing hypotheses about locomotion. Perceived risk (or, the “landscape of fear”) is a key pressure that has selected for primate arboreality – particularly nocturnal arboreality. We propose that shifts in Plio-Pleistocene landscapes of fear – caused by declining carnivoran abundance and diversity – might also have been a key selection pressure in changes to primate locomotion, particularly papionin and hominid terrestriality. We discuss this hypothesis and propose future research avenues to explore it further. Not only will such research provide a more nuanced view of the causes and consequences of a rare behavioural trait in primates, but it could ultimately help us explain how one group of African apes came to spend all their time on the ground, and how that made them human.
During excavations of historical monuments, archaeologists find various artifacts that testify to the existence and everyday life of our distant ancestors. Particular attention is paid to the remains of living organisms. They not only provide evidence of the economic activity of ancient farmers, but also help to identify phylogenetic relationships and domestication processes in the world's centers of diversity. Due to the long-term presence of paleontological objects in the environment that is not conducive to preservation, they often get destroyed and it becomes impossible to determine which species they belong to. Therefore, archaeologists increasingly resort to the help of paleogeneticists. The works on studies of ancient DNA (aDNA) from human and animal remains are known in Russia. However, paleogenetic studies of fossil plant remains such as pollen, seeds, and timber are few. In 2019, carbonized grains of cereal crops were found on the territory of the Usvyaty settlement in Pskov Region. The findings date back to the 12th century. The morphological analysis of the seed mixture resulted in finding grains, the degree of destruction of which prevented determination of the species they belong to by analyzing their microrelief. Therefore, the aim of this study was to develop taxon-specific primers that yield a short amplification product for the analysis of fragmented aDNA from the destroyed barley caryopses. As a result, a PCR test named HORDELF was developed, which is recommended for the identification of plant residues (carbonized seeds) belonging to the genus Hordeum L.
Paulo Ricardo de Oliveira Costa, A. Chahud, Mercedes Okumura
The Ribeira of Iguape Valley, located in the south of São Paulo State, presents an important cave system where several preserved Quaternary faunal remains have been found, especially in the Abismo Ponta de Flecha, Abismo do Fóssil and Abismo do Juvenal caves. Although the records of toxodontids are common in the region, almost no multidisciplinary research has been done in these materials, including the potential interaction with prehistoric humans. Therefore, the objective of this work was to analyze teeth of toxodontids found in the Ribeira Valley, to understand paleontological aspects of these animals and investigate the presence of cut marks as a proxy for potential interactions with human groups. Taxonomic analysis and measurements showed that all teeth could be assigned to the species Toxodon platensis, deciduous teeth to the cf. T. platensis, and one lower molar that was suggestively associated with Mixotoxodon larensis. The analysis of paleopathologies identified the presence of dental enamel hypoplasia in all specimens found at Abismo Ponta de Flecha Cave, suggesting that this individual suffered from stress events during the period of tooth formation. Cut marks were identified in only two specimens of T. platensis, but the unusual location of these incisions and the identified weathering marks made it difficult to understand the aiming of these marks. Keywords: enamel hypoplasia, human marks, Mixotoxodon, Toxodon.
Bernat Josep Vázquez López, Albert Sellés, Albert Prieto-Márquez
et al.
Abstract Paleoenvironmental preferences for Cretaceous dinosaurs at a regional scale have been mainly assessed in North America. In south-western Europe, the dinosaur-bearing formations ranging the late Campanian to the latest Maastrichtian encompass coastal and lowland environments that produced hundreds of fossil localities with evidence of titanosaurian sauropods, maniraptoran and abelisauroid theropods, and nodosaurid ankylosaurs, together with rhabdodontid and hadrosauroid ornithopods. In order to study environmental associations of dinosaur taxa, we have revised, updated, and expanded upon an existing database that compiles the occurrence and minimum number of individuals for the dinosaur-bearing formations spanning the upper Campanian to the uppermost Maastrichtian of South-Western Europe. Based on this database, the habitat preferences of dinosaur groups in the region were determined by means of statistical tests of independence. All chi-square tests showed positive, mostly moderate-to-strong, and statistically significant associations between the studied groups and the environment they inhabited. The analysis of the residuals indicated that most dinosaur groups preferred lowland environments (including, contrary to previous studies, nodosaurids). The only exception were abelisauroids, which showed no habitat preference. Our results concur with recent works indicating that titanosaur sauropods and hadrosauroids preferred inland environments but clearly disagree with others suggesting that the latter as well as nodosaurid ankylosaurs were positively associated with marine or coastal settings. Considering the changes in occurrence distribution throughout the Maastrichtian turnover in the region, both titanosaurians and nodosaurids probably stablished a feeding strategy-based niche partitioning with ornithopods, although additional data is required to confidently confirm this relationship.
Visual systems have been crucial for animals to detect light signals. Binocular stereopsis has affected prey-predator re
lationships throughout animal evolution by providing depth perception, among others. However, it has been difficult to
reconstruct extinct binocular functions due to a lack of suitable fossil material. Here, we show, based on morphological
analysis of well-preserved eyes, that an extinct mantis (Ambermantis wozniaki Grimaldi, 2003) in the Cretaceous New
Jersey amber developed an advanced visual system as a predator. We found that A. wozniaki possesses large compound
eyes with numerous, ca. 12 000 ommatidia. The interocular distance is narrower than the eyes, and the estimated bin
ocular visual field is broader than in the typical extant basal and derived taxa. The large number of ommatidia indicates
that the compound eyes of A. wozniaki achieved high spatial resolution to capture objects visually. The broad binocular
field supports that A. wozniaki increased the stereoscopic area and developed an advanced prey-recognition system.
These findings suggest that the Cretaceous basal mantises were highly adaptive visual predators, implying the ecological
domination of mantises as visual specialists for 90 million years.
Isolated perspectives have often paved the way for great scientific discoveries. However, many breakthroughs only emerged when moving away from singular views towards interactions. Discussions on Artificial Intelligence (AI) typically treat human and AI bias as distinct challenges, leaving their dynamic interplay and compounding potential largely unexplored. Recent research suggests that biased AI can amplify human cognitive biases, while well-calibrated systems might help mitigate them. In this position paper, I advocate for transcending beyond separate treatment of human and AI biases and instead focus on their interaction effects. I argue that a comprehensive framework, one that maps (compound human-AI) biases to mitigation strategies, is essential for understanding and protecting human cognition, and I outline concrete steps for its development.
Tailia Malloy, Maria Jose Ferreira, Fei Fang
et al.
In real-world decision making, outcomes are often delayed, meaning individuals must make multiple decisions before receiving any feedback. Moreover, feedback can be presented in different ways: it may summarize the overall results of multiple decisions (aggregated feedback) or report the outcome of individual decisions after some delay (clustered feedback). Despite its importance, the timing and presentation of delayed feedback has received little attention in cognitive modeling of decision-making, which typically focuses on immediate feedback. To address this, we conducted an experiment to compare the effect of delayed vs. immediate feedback and aggregated vs. clustered feedback. We also propose a Hierarchical Instance-Based Learning (HIBL) model that captures how people make decisions in delayed feedback settings. HIBL uses a super-model that chooses between sub-models to perform the decision-making task until an outcome is observed. Simulations show that HIBL best predicts human behavior and specific patterns, demonstrating the flexibility of IBL models.
The gaze of a person tends to reflect their interest. This work explores what happens when this statement is taken literally and applied to robots. Here we present a robot system that employs a moving robot head with a screen-based eye model that can direct the robot's gaze to points in physical space and present a reflection-like mirror image of the attended region on top of each eye. We conducted a user study with 33 participants, who were asked to instruct the robot to perform pick-and-place tasks, monitor the robot's task execution, and interrupt it in case of erroneous actions. Despite a deliberate lack of instructions about the role of the eyes and a very brief system exposure, participants felt more aware about the robot's information processing, detected erroneous actions earlier, and rated the user experience higher when eye-based mirroring was enabled compared to non-reflective eyes. These results suggest a beneficial and intuitive utilization of the introduced method in cooperative human-robot interaction.
High-stakes decision domains are increasingly exploring the potential of Large Language Models (LLMs) for complex decision-making tasks. However, LLM deployment in real-world settings presents challenges in data security, evaluation of its capabilities outside controlled environments, and accountability attribution in the event of adversarial decisions. This paper proposes a framework for responsible deployment of LLM-based decision-support systems through active human involvement. It integrates interactive collaboration between human experts and developers through multiple iterations at the pre-deployment stage to assess the uncertain samples and judge the stability of the explanation provided by post-hoc XAI techniques. Local LLM deployment within organizations and decentralized technologies, such as Blockchain and IPFS, are proposed to create immutable records of LLM activities for automated auditing to enhance security and trace back accountability. It was tested on Bert-large-uncased, Mistral, and LLaMA 2 and 3 models to assess the capability to support responsible financial decisions on business lending.
Eleni Straitouri, Stratis Tsirtsis, Ander Artola Velasco
et al.
Recent work has shown that, in classification tasks, it is possible to design decision support systems that do not require human experts to understand when to cede agency to a classifier or when to exercise their own agency to achieve complementarity$\unicode{x2014}$experts using these systems make more accurate predictions than those made by the experts or the classifier alone. The key principle underpinning these systems reduces to adaptively controlling the level of human agency, by design. Can we use the same principle to achieve complementarity in sequential decision making tasks? In this paper, we answer this question affirmatively. We develop a decision support system that uses a pre-trained AI agent to narrow down the set of actions a human can take to a subset, and then asks the human to take an action from this action set. Along the way, we also introduce a bandit algorithm that leverages the smoothness properties of the action sets provided by our system to efficiently optimize the level of human agency. To evaluate our decision support system, we conduct a large-scale human subject study ($n = 1{,}600$) where participants play a wildfire mitigation game. We find that participants who play the game supported by our system outperform those who play on their own by $\sim$$30$% and the AI agent used by our system by $>$$2$%, even though the AI agent largely outperforms participants playing without support. We have made available the data gathered in our human subject study as well as an open source implementation of our system at https://github.com/Networks-Learning/narrowing-action-choices .
Traditional image annotation tasks rely heavily on human effort for object selection and label assignment, making the process time-consuming and prone to decreased efficiency as annotators experience fatigue after extensive work. This paper introduces a novel framework that leverages the visual understanding capabilities of large multimodal models (LMMs), particularly GPT, to assist annotation workflows. In our proposed approach, human annotators focus on selecting objects via bounding boxes, while the LMM autonomously generates relevant labels. This human-AI collaborative framework enhances annotation efficiency by reducing the cognitive and time burden on human annotators. By analyzing the system's performance across various types of annotation tasks, we demonstrate its ability to generalize to tasks such as object recognition, scene description, and fine-grained categorization. Our proposed framework highlights the potential of this approach to redefine annotation workflows, offering a scalable and efficient solution for large-scale data labeling in computer vision. Finally, we discuss how integrating LMMs into the annotation pipeline can advance bidirectional human-AI alignment, as well as the challenges of alleviating the "endless annotation" burden in the face of information overload by shifting some of the work to AI.
An inventory of the paleontological resources in Virginia caves was initiated by the authors in 1996. Invertebrate and vertebrate fossil occurrences in Virginia were examined to understand the range and scope of fossils present in caves to aid in determination as to what should constitute cave significance in the documentation of paleontological resources for Virginia’s Significant Cave List. This cave study reviewed previously known vertebrate fossil sites and new vertebrate and invertebrate fossil sites before focusing on vertebrate fossil occurrences and the evidence of how they came to be deposited. As part of the inventory, to address the question of the percentage of Virginia caves that are paleontologically significant, 229 Virginia caves were examined to identify 100 caves with exposed non-human bones and/or teeth. Thirteen of the 100 caves yielded extinct or significant extirpated taxa, including one determined to have been excavated by E.D. Cope in 1867. Thirteen percent of the 100 bone caves examined or 5.7 % of the 229 study caves were documented as paleontologically significant. A selection of vertebrate cave fossils were radiocarbon dated and ranged in age from 11,986 +/- 76 to 39,300 +/- 1100 YBP. Uranium-thorium dating at one site suggests the fossil were older than 300,000 years and may be older than 450,000 YBP, indicating the age of this bone deposit is Pleistocene (Irvingtonian). A total of 43 cave sites were documented as paleontologically significant in the inventory and reported in Appendix C. PRIOVAC was recognized as an NSS Project from 2001 through 2006 and progress talks were presented at NSS Conventions in 1995, 1997, 1998, 2000, 2001, 2002, 2006, 2007, and a version of this summary in 2023 with abstracts published in the NSS Journal of Cave and Karst Studies and talks at the UIS Congresses in 1997 (Switzerland), 2001 (Brazil), and 2009 (US) published as papers in the UIS Proceedings. This paper is a summary report on PRIOVAC and the first publication of some of the bone dates and the U-Th dates.
Mariel Ferrari, Verónica Bertero, Marcelo G. Carrera
Gastropods from the Upper Ordovician of the Argentine Precordillera received less attention than other coeval marine invertebrates in this region. The present contribution supplies accurate taxonomic information recovering 10 gastropod genera which are represented by 10 species from the La Pola and Don Braulio formations (Sandbian and Hirnantian units, respectively) at the San Juan Province, Argentina: two species, namely Tetranota argentina sp. nov. and Clathrospira gondwanica sp. nov., are new to science. The gastropod association shows a remarkable diversity, and in contrast to other invertebrate groups reported in the Argentine Precordillera (e.g., trilobites, ostracodes, brachiopods, sponges, and bivalves) which had Gondwanan affinities during the Late Ordovician, the gastropod assemblage from the La Pola and Don Braulio formations had major palaeobiogeographical similarities with their Northern Hemisphere (Laurentia, Avalonia, and Baltica) counterparts. The occurrence of the genus Clathrospira suggests that it could have been the first precursor of the order Pleurotomariida in South America during the Sandbian–Hirnantian, and can help to shed light on the origin of this clade in the southernmost Gondwana continent as early as Palaeozoic times.
People often use visualizations not only to explore a dataset but also to draw generalizable conclusions about underlying models or phenomena. While previous research has viewed deviations from rational analysis as problematic, we hypothesize that human reliance on non-normative heuristics may be advantageous in certain situations. In this study, we investigate scenarios where human intuition might outperform idealized statistical rationality. Our experiment assesses participants' accuracy in characterizing the parameters of known data-generating models from bivariate visualizations. Our findings show that, while participants generally demonstrated lower accuracy than statistical models, they often outperformed Bayesian agents, particularly when dealing with extreme samples. These results suggest that, even when deviating from rationality, human gut reactions to visualizations can provide an advantage. Our findings offer insights into how analyst intuition and statistical models can be integrated to improve inference and decision-making, with important implications for the design of visual analytics tools.
Homelessness is a humanitarian challenge affecting an estimated 1.6 billion people worldwide. In the face of rising homeless populations in developed nations and a strain on social services, government agencies are increasingly adopting data-driven models to determine one's risk of experiencing homelessness and assigning scarce resources to those in need. We conducted a systematic literature review of 57 papers to understand the evolution of these decision-making algorithms. We investigated trends in computational methods, predictor variables, and target outcomes used to develop the models using a human-centered lens and found that only 9 papers (15.7%) investigated model fairness and bias. We uncovered tensions between explainability and ecological validity wherein predictive risk models (53.4%) focused on reductive explainability while resource allocation models (25.9%) were dependent on unrealistic assumptions and simulated data that are not useful in practice. Further, we discuss research challenges and opportunities for developing human-centered algorithms in this area.
Children with Autism Spectrum Disorder (ASD) often exhibit atypical facial expressions. However, the specific objective facial features that underlie this subjective perception remain unclear. In this paper, we introduce a novel dataset, Hugging Rain Man (HRM), which includes facial action units (AUs) manually annotated by FACS experts for both children with ASD and typical development (TD). The dataset comprises a rich collection of posed and spontaneous facial expressions, totaling approximately 130,000 frames, along with 22 AUs, 10 Action Descriptors (ADs), and atypicality ratings. A statistical analysis of static images from the HRM reveals significant differences between the ASD and TD groups across multiple AUs and ADs when displaying the same emotional expressions, confirming that participants with ASD tend to demonstrate more irregular and diverse expression patterns. Subsequently, a temporal regression method was presented to analyze atypicality of dynamic sequences, thereby bridging the gap between subjective perception and objective facial characteristics. Furthermore, baseline results for AU detection are provided for future research reference. This work not only contributes to our understanding of the unique facial expression characteristics associated with ASD but also provides potential tools for ASD early screening. Portions of the dataset, features, and pretrained models are accessible at: \url{https://github.com/Jonas-DL/Hugging-Rain-Man}.
Paula Akemi Aoyagui, Sharon Ferguson, Anastasia Kuzminykh
An essential aspect of evaluating Large Language Models (LLMs) is identifying potential biases. This is especially relevant considering the substantial evidence that LLMs can replicate human social biases in their text outputs and further influence stakeholders, potentially amplifying harm to already marginalized individuals and communities. Therefore, recent efforts in bias detection invested in automated benchmarks and objective metrics such as accuracy (i.e., an LLMs output is compared against a predefined ground truth). Nonetheless, social biases can be nuanced, oftentimes subjective and context-dependent, where a situation is open to interpretation and there is no ground truth. While these situations can be difficult for automated evaluation systems to identify, human evaluators could potentially pick up on these nuances. In this paper, we discuss the role of human evaluation and subjective interpretation to augment automated processes when identifying biases in LLMs as part of a human-centred approach to evaluate these models.