Search

Large Multimodal Models (LMMs) exhibit major shortfalls when interpreting images and, by some measures, have poorer spatial cognition than small children or animals. Despite this, they attain high scores on many popular visual benchmarks, with headroom rapidly eroded by an ongoing surge of model progress. To address this, there is a pressing need for difficult benchmarks that remain relevant for longer. We take this idea to its limit by introducing ZeroBench-a lightweight visual reasoning benchmark that is entirely impossible for contemporary frontier LMMs. Our benchmark consists of 100 manually curated questions and 334 less difficult subquestions. We evaluate 20 LMMs on ZeroBench, all of which score 0.0%, and rigorously analyse the errors. To encourage progress in visual understanding, we publicly release ZeroBench., Comment: 20 pages, 13 figures

Bayesian neural networks (BNNs), which estimate the full posterior distribution over model parameters, are well-known for their role in uncertainty quantification and its promising application in out-of-distribution detection (OoD). Amongst other uncertainty measures, BNNs provide a state-of-the art estimation of predictive entropy (total uncertainty) which can be decomposed as the sum of mutual information and expected entropy. In the context of OoD detection the estimation of predictive uncertainty in the form of the predictive entropy score confounds aleatoric and epistemic uncertainty, the latter being hypothesized to be high for OoD points. Despite these justifications, the mutual information score has been shown to perform worse than predictive entropy. Taking inspiration from Bayesian variational autoencoder (BVAE) literature, this work proposes to measure the disagreement between a corrected version of the pre-softmax quantities, otherwise known as logits, as an estimate of epistemic uncertainty for Bayesian NNs under mean field variational inference. The three proposed epistemic uncertainty scores demonstrate marked improvements over mutual information on a range of OoD experiments, with equal performance otherwise. Moreover, the epistemic uncertainty scores perform on par with the Bayesian benchmark predictive entropy on a range of MNIST and CIFAR10 experiments., Comment: Presented at ECCV 2024 Workshop: 3rd Workshop on Uncertainty Quantification for Computer Vision

Transmission Control Protocol (TCP) continues to be the dominant transport protocol on the Internet. The stability of fluid models has been a key consideration in the design of TCP and the performance evaluation of TCP algorithms. Based on local stability analysis, we formulate some design considerations for a class of TCP algorithms. We begin with deriving sufficient conditions for the local stability of a generalized TCP algorithm in the presence of heterogeneous round-trip delays. Within this generalized model, we consider three specific variants of TCP: TCP Reno, Compound TCP, and Scalable TCP. The sufficient conditions we derive are scalable across network topologies with one, two, and many bottleneck links. We are interested in networks with intermediate and small drop-tail buffers as they offer smaller queuing delays. The small buffer regime is more attractive as the conditions for stability are decentralized. TCP algorithms that follow our design considerations can provide stable operation on any network topology, irrespective of the number of bottleneck links or delays in the network.

We present PrismAvatar: a 3D head avatar model which is designed specifically to enable real-time animation and rendering on resource-constrained edge devices, while still enjoying the benefits of neural volumetric rendering at training time. By integrating a rigged prism lattice with a 3D morphable head model, we use a hybrid rendering model to simultaneously reconstruct a mesh-based head and a deformable NeRF model for regions not represented by the 3DMM. We then distill the deformable NeRF into a rigged mesh and neural textures, which can be animated and rendered efficiently within the constraints of the traditional triangle rendering pipeline. In addition to running at 60 fps with low memory usage on mobile devices, we find that our trained models have comparable quality to state-of-the-art 3D avatar models on desktop devices., Comment: 8 pages, 5 figures

Lateral Spine Image (LSI) analysis is important for medical diagnosis, treatment planning, and detailed spinal health assessments. Although modalities like Computed Tomography and Digital X-ray Imaging are commonly used, Dual Energy X-ray Absorptiometry (DXA) is often preferred due to lower radiation exposure, seamless capture, and cost-effectiveness. Accurate Vertebral Landmark Localization (VLL) on LSIs is important to detect spinal conditions like kyphosis and lordosis, as well as assessing Abdominal Aortic Calcification (AAC) using Inter-Vertebral Guides (IVGs). Nonetheless, few automated VLL methodologies have concentrated on DXA LSIs. We present VerteNet, a hybrid CNN-Transformer model featuring a novel dual-resolution attention mechanism in self and cross-attention domains, referred to as Dual Resolution Self-Attention (DRSA) and Dual Resolution Cross-Attention (DRCA). These mechanisms capture the diverse frequencies in DXA images by operating at two different feature map resolutions. Additionally, we design a Multi-Context Feature Fusion Block (MCFB) that efficiently integrates the features using DRSA and DRCA. We train VerteNet on 620 DXA LSIs from various machines and achieve superior results compared to existing methods. We also design an algorithm that utilizes VerteNet's predictions in estimating the Region of Interest (ROI) to detect potential abdominal aorta cropping, where inadequate soft tissue hinders calcification assessment. Additionally, we present a small proof-of-concept study to show that IVGs generated from VLL information can improve inter-reader correlation in AAC scoring, addressing two key areas of disagreement in expert AAC-24 scoring: IVG placement and quality control for full abdominal aorta assessment. The code for this work can be found at https://github.com/zaidilyas89/VerteNet., Comment: 10 pages with 7 figures

The last few years have witnessed a spate of data protection regulations in conjunction with an ever-growing appetite for data usage in large businesses, thus presenting significant challenges for businesses to maintain compliance. To address this conflict, we present Data Guard - a fine-grained, purpose-based access control system for large data warehouses. Data Guard enables authoring policies based on semantic descriptions of data and purpose of data access. Data Guard then translates these policies into SQL views that mask data from the underlying warehouse tables. At access time, Data Guard ensures compliance by transparently routing each table access to the appropriate data-masking view based on the purpose of the access, thus minimizing the effort of adopting Data Guard in existing applications. Our enforcement solution allows masking data at much finer granularities than what traditional solutions allow. In addition to row and column level data masking, Data Guard can mask data at the sub-cell level for columns with non-atomic data types such as structs, arrays, and maps. This fine-grained masking allows Data Guard to preserve data utility for consumers while ensuring compliance. We implemented a number of performance optimizations to minimize the overhead of data masking operations. We perform numerous experiments to identify the key factors that influence the data masking overhead and demonstrate the efficiency of our implementation.

We construct and analyze the fault tolerance of $[[6,1,3]]$ non-CSS quantum error correcting code under the anisotropic and depolarizing noise models. This rate-optimized code achieves fault-tolerance using a single ancilla qubit for syndrome measurement under anisotropic noise conditions. This method was called fault-tolerance using bare ancilla by Brown \emph{et al.} We give explicit construction of the code using measurements on non-planar graph states. We also argue that using our approach, we can construct a family of such fault-tolerant codes. This method fills a notable gap in constructing fault-tolerant non-CSS code families., Comment: 10 pages, 12 figures

The fault tolerance of quantum devices requires on-par contributions from error-correcting codes and suitable decoders. One of the most explored error-correcting codes is the family of Quantum Low-Density Parity Check (QLDPC) codes. Although faster than many of the reported decoders for QLDPC codes, iterative decoders fail due to the colossal degeneracy and short cycles intrinsic to these codes. We present a strategy to improve the performance of the iterative decoders based on a collaborative way to use the message passing of the iterative decoders and check node removal from the code's Tanner graph. We use the concept of bit separation and generalize it to qubit separation. This gives us a metric to analyze and improve the decoder's performance towards harmful configurations of QLDPC codes. We present a simple decoding architecture to overcome iterative decoding failures by increasing the separation of trapped qubits without incurring any significant overhead., Comment: 11 pages, 6 figures

Quantum simulation in its current state faces experimental overhead in terms of physical space and cooling. We propose boson sampling as an alternative compact synthetic platform performing at room temperature. Identifying the capability of estimating matrix permanents, we explore the applicability of boson sampling for tackling the dynamics of quantum systems without having access to information about the full state vector. By mapping the time-evolution unitary of a Hamiltonian onto an interferometer via continuous-variable gate decompositions, we present proof-of-principle results of localization characteristics of a single particle. We study the dynamics of one-dimensional tight-binding systems in the clean and quasiperiodic-disordered limits to observe Bloch oscillations and dynamical localization, and the delocalization-to-localization phase transition in the Aubry- Andre-Harper model respectively. Our computational results obtained using boson sampling are in complete agreement with the dynamical and static results of non-interacting tight-binding systems obtained using conventional numerical calculations. Additionally, our study highlights the role of number of sampling measurements or shots for simulation accuracy., Comment: 14 pages, 11 figures

Artificial Intelligence (AI) has witnessed major growth and integration across various domains. As AI systems become increasingly prevalent, they also become targets for threat actors to manipulate their functionality for malicious purposes. This survey paper on offensive AI will comprehensively cover various aspects related to attacks against and using AI systems. It will delve into the impact of offensive AI practices on different domains, including consumer, enterprise, and public digital infrastructure. The paper will explore adversarial machine learning, attacks against AI models, infrastructure, and interfaces, along with offensive techniques like information gathering, social engineering, and weaponized AI. Additionally, it will discuss the consequences and implications of offensive AI, presenting case studies, insights, and avenues for further research.

Automated assessment in natural language generation is a challenging task. Instruction-tuned large language models (LLMs) have shown promise in reference-free evaluation, particularly through comparative assessment. However, the quadratic computational complexity of pairwise comparisons limits its scalability. To address this, efficient comparative assessment has been explored by applying comparative strategies on zero-shot LLM probabilities. We propose a framework for finetuning LLMs for comparative assessment to align the model's output with the target distribution of comparative probabilities. By training on soft probabilities, our approach improves state-of-the-art performance while maintaining high performance with an efficient subset of comparisons., Comment: 8 pages, 5 figures, 6 tables

We consider the Weak Galerkin finite element approximation of the Singularly Perturbed Biharmonic elliptic problem on a unit square domain with clamped boundary conditions. Shishkin mesh is used for domain discretization as the solution exhibits boundary layers near the domain boundary. Error estimates in the equivalent $H^{2}-$ norm have been established and the uniform convergence of the proposed method has been proved. Numerical examples are presented corroborating our theoretical findings.

Ultrasound is widely employed for clinical intervention and diagnosis, due to its advantages of offering non-invasive, radiation-free, and real-time imaging. However, the accessibility of this dexterous procedure is limited due to the substantial training and expertise required of operators. The robotic ultrasound (RUS) offers a viable solution to address this limitation; nonetheless, achieving human-level proficiency remains challenging. Learning from demonstrations (LfD) methods have been explored in RUS, which learns the policy prior from a dataset of offline demonstrations to encode the mental model of the expert sonographer. However, active engagement of experts, i.e. Coaching, during the training of RUS has not been explored thus far. Coaching is known for enhancing efficiency and performance in human training. This paper proposes a coaching framework for RUS to amplify its performance. The framework combines DRL (self-supervised practice) with sparse expert's feedback through coaching. The DRL employs an off-policy Soft Actor-Critic (SAC) network, with a reward based on image quality rating. The coaching by experts is modeled as a Partially Observable Markov Decision Process (POMDP), which updates the policy parameters based on the correction by the expert. The validation study on phantoms showed that coaching increases the learning rate by $25\%$ and the number of high-quality image acquisition by $74.5\%$., Comment: Accepted in IEEE Transactions on Medical Robotics and Bionics (TMRB) 2024

This technical report describes the implementation of Toychain: a simple, lightweight blockchain implemented in Python, designed for ease of deployment and practicality in robotics research. It can be integrated with various software and simulation tools used in robotics (we have integrated it with ARGoS, Gazebo, and ROS2), and also be deployed on real robots capable of Wi-Fi communications. The Toychain package supports the deployment of smart contracts written in Python (computer programs that can be executed by and synchronized across a distributed network). The nodes in the blockchain can execute smart contract functions by broadcasting transactions, which update the state of the blockchain upon agreement by all other nodes. The conditions for this agreement are established by a consensus protocol. The Toychain package allows for custom implementations of the consensus protocol, which can be useful for research or meeting specific application requirements. Currently, Proof-of-Work and Proof-of-Authority are implemented.

Speech enabled foundation models, either in the form of flexible speech recognition based systems or audio-prompted large language models (LLMs), are becoming increasingly popular. One of the interesting aspects of these models is their ability to perform tasks other than automatic speech recognition (ASR) using an appropriate prompt. For example, the OpenAI Whisper model can perform both speech transcription and speech translation. With the development of audio-prompted LLMs there is the potential for even greater control options. In this work we demonstrate that with this greater flexibility the systems can be susceptible to model-control adversarial attacks. Without any access to the model prompt it is possible to modify the behaviour of the system by appropriately changing the audio input. To illustrate this risk, we demonstrate that it is possible to prepend a short universal adversarial acoustic segment to any input speech signal to override the prompt setting of an ASR foundation model. Specifically, we successfully use a universal adversarial acoustic segment to control Whisper to always perform speech translation, despite being set to perform speech transcription. Overall, this work demonstrates a new form of adversarial attack on multi-tasking speech enabled foundation models that needs to be considered prior to the deployment of this form of model.

Telerobotic and Autonomous Robotic Ultrasound Systems (RUS) help alleviate the need for operator-dependability in free-hand ultrasound examinations. However, the state-of-the-art RUSs still rely on a human operator to apply the ultrasound gel. The lack of standardization in this process often leads to poor imaging of the scanned region. The reason for this has to do with air-gaps between the probe and the human body. In this paper, we developed a end-of-arm tool for RUS, referred to as UltraGelBot. This bot can autonomously detect and dispense the gel. It uses a deep learning model to detect the gel from images acquired using an on-board camera. A motorized mechanism is also developed, which will use this feedback and dispense the gel. Experiments on phantom revealed that UltraGelBot increases the acquired image quality by $18.6\%$ and reduces the procedure time by $37.2\%$., Comment: 2024 16th Hamlyn Symposium on Medical Robotics (HSMR)

Enterprise retrieval augmented generation (RAG) offers a highly flexible framework for combining powerful large language models (LLMs) with internal, possibly temporally changing, documents. In RAG, documents are first chunked. Relevant chunks are then retrieved for a user query, which are passed as context to a synthesizer LLM to generate the query response. However, the retrieval step can limit performance, as incorrect chunks can lead the synthesizer LLM to generate a false response. This work applies a zero-shot adaptation of standard dense retrieval steps for more accurate chunk recall. Specifically, a chunk is first decomposed into atomic statements. A set of synthetic questions are then generated on these atoms (with the chunk as the context). Dense retrieval involves finding the closest set of synthetic questions, and associated chunks, to the user query. It is found that retrieval with the atoms leads to higher recall than retrieval with chunks. Further performance gain is observed with retrieval using the synthetic questions generated over the atoms. Higher recall at the retrieval step enables higher performance of the enterprise LLM using the RAG pipeline., Comment: 14 pages, 5 figures, 5 tables

Recent developments in large speech foundation models like Whisper have led to their widespread use in many automatic speech recognition (ASR) applications. These systems incorporate `special tokens' in their vocabulary, such as $\texttt{<|endoftext|>}$, to guide their language generation process. However, we demonstrate that these tokens can be exploited by adversarial attacks to manipulate the model's behavior. We propose a simple yet effective method to learn a universal acoustic realization of Whisper's $\texttt{<|endoftext|>}$ token, which, when prepended to any speech signal, encourages the model to ignore the speech and only transcribe the special token, effectively `muting' the model. Our experiments demonstrate that the same, universal 0.64-second adversarial audio segment can successfully mute a target Whisper ASR model for over 97\% of speech samples. Moreover, we find that this universal adversarial audio segment often transfers to new datasets and tasks. Overall this work demonstrates the vulnerability of Whisper models to `muting' adversarial attacks, where such attacks can pose both risks and potential benefits in real-world settings: for example the attack can be used to bypass speech moderation systems, or conversely the attack can also be used to protect private speech data.

Compiling programs to an instruction set architecture (ISA) requires a set of rewrite rules that map patterns consisting of compiler instructions to patterns consisting of ISA instructions. We synthesize such rules by constructing SMT queries, whose solutions represent two functionally equivalent programs. These two programs are interpreted as an instruction selection rewrite rule. Existing work is limited to single-instruction ISA patterns, whereas our solution does not have that restriction. Furthermore, we address inefficiencies of existing work by developing two optimized algorithms. The first only generates unique rules by preventing synthesis of duplicate and composite rules. The second only generates lowest-cost rules by preventing synthesis of higher-cost rules. We evaluate our algorithms on multiple ISAs. Without our optimizations, the vast majority of synthesized rewrite rules are either duplicates, composites, or higher cost. Our optimizations result in synthesis speed-ups of up to 768x and 4004x for the two algorithms., Comment: 9 pages, 2 figures, in submission to FMCAD24, fixed small bugs in results generation, improved baseline comparison, improved background and writing clarity

LLM-as-a-judge approaches are a practical and effective way of assessing a range of text tasks. However, when using pairwise comparisons to rank a set of candidates, the computational cost scales quadratically with the number of candidates, which has practical limitations. This paper introduces a Product of Expert (PoE) framework for efficient LLM Comparative Assessment. Here individual comparisons are considered experts that provide information on a pair's score difference. The PoE framework combines the information from these experts to yield an expression that can be maximized with respect to the underlying set of candidates, and is highly flexible where any form of expert can be assumed. When Gaussian experts are used one can derive simple closed-form solutions for the optimal candidate ranking, and expressions for selecting which comparisons should be made to maximize the probability of this ranking. Our approach enables efficient comparative assessment, where by using only a small subset of the possible comparisons, one can generate score predictions that correlate well with human judgements. We evaluate the approach on multiple NLG tasks and demonstrate that our framework can yield considerable computational savings when performing pairwise comparative assessment. With many candidate texts, using as few as 2% of comparisons the PoE solution can achieve similar performance to when all comparisons are used.

Multiple-choice (MC) tests are an efficient method to assess English learners. It is useful for test creators to rank candidate MC questions by difficulty during exam curation. Typically, the difficulty is determined by having human test takers trial the questions in a pretesting stage. However, this is expensive and not scalable. Therefore, we explore automated approaches to rank MC questions by difficulty. However, there is limited data for explicit training of a system for difficulty scores. Hence, we compare task transfer and zero-shot approaches: task transfer adapts level classification and reading comprehension systems for difficulty ranking while zero-shot prompting of instruction finetuned language models contrasts absolute assessment against comparative. It is found that level classification transfers better than reading comprehension. Additionally, zero-shot comparative assessment is more effective at difficulty ranking than the absolute assessment and even the task transfer approaches at question difficulty ranking with a Spearman's correlation of 40.4%. Combining the systems is observed to further boost the correlation., Comment: 7 pages, 3 figures

The mechanisms of psychotic symptoms like hallucinations and delusions are often investigated in fully-formed illness, well after symptoms emerge. These investigations have yielded key insights, but are not well-positioned to reveal the dynamic forces underlying symptom formation itself. Understanding symptom development over time would allow us to identify steps in the pathophysiological process leading to psychosis, shifting the focus of psychiatric intervention from symptom alleviation to prevention. We propose a model for understanding the emergence of psychotic symptoms within the context of an adaptive, developing neural system. We will make the case for a pathophysiological process that begins with cortical hyperexcitability and bottom-up noise transmission, which engenders inappropriate belief formation via aberrant prediction error signaling. We will argue that this bottom-up noise drives learning about the (im)precision of new incoming sensory information because of diminished signal-to-noise ratio, causing an adaptive relative over-reliance on prior beliefs. This over-reliance on priors predisposes to hallucinations and covaries with hallucination severity. An over-reliance on priors may also lead to increased conviction in the beliefs generated by bottom-up noise and drive movement toward conversion to psychosis. We will identify predictions of our model at each stage, examine evidence to support or refute those predictions, and propose experiments that could falsify or help select between alternative elements of the overall model. Nesting computational abnormalities within longitudinal development allows us to account for hidden dynamics among the mechanisms driving symptom formation and to view established symptomatology as a point of equilibrium among competing biological forces.

When working with 3D facial data, improving fidelity and avoiding the uncanny valley effect is critically dependent on accurate 3D facial performance capture. Because such methods are expensive and due to the widespread availability of 2D videos, recent methods have focused on how to perform monocular 3D face tracking. However, these methods often fall short in capturing precise facial movements due to limitations in their network architecture, training, and evaluation processes. Addressing these challenges, we propose a novel face tracker, FlowFace, that introduces an innovative 2D alignment network for dense per-vertex alignment. Unlike prior work, FlowFace is trained on high-quality 3D scan annotations rather than weak supervision or synthetic data. Our 3D model fitting module jointly fits a 3D face model from one or many observations, integrating existing neutral shape priors for enhanced identity and expression disentanglement and per-vertex deformations for detailed facial feature reconstruction. Additionally, we propose a novel metric and benchmark for assessing tracking accuracy. Our method exhibits superior performance on both custom and publicly available benchmarks. We further validate the effectiveness of our tracker by generating high-quality 3D data from 2D videos, which leads to performance gains on downstream tasks., Comment: 22 pages, 25 figures, to be published in CVPR 2024

This study examines the use of large language models (LLMs) by undergraduate and graduate students for programming assignments in advanced computing classes. Unlike existing research, which primarily focuses on introductory classes and lacks in-depth analysis of actual student-LLM interactions, our work fills this gap. We conducted a comprehensive analysis involving 411 students from a Distributed Systems class at an Indian university, where they completed three programming assignments and shared their experiences through Google Form surveys. Our findings reveal that students leveraged LLMs for a variety of tasks, including code generation, debugging, conceptual inquiries, and test case creation. They employed a spectrum of prompting strategies, ranging from basic contextual prompts to advanced techniques like chain-of-thought prompting and iterative refinement. While students generally viewed LLMs as beneficial for enhancing productivity and learning, we noted a concerning trend of over-reliance, with many students submitting entire assignment descriptions to obtain complete solutions. Given the increasing use of LLMs in the software industry, our study highlights the need to update undergraduate curricula to include training on effective prompting strategies and to raise awareness about the benefits and potential drawbacks of LLM usage in academic settings., Comment: Under review: 8 pages