Search

We introduce RNA-FrameFlow, the first generative model for 3D RNA backbone design. We build upon SE(3) flow matching for protein backbone generation and establish protocols for data preparation and evaluation to address unique challenges posed by RNA modeling. We formulate RNA structures as a set of rigid-body frames and associated loss functions which account for larger, more conformationally flexible RNA backbones (13 atoms per nucleotide) vs. proteins (4 atoms per residue). Toward tackling the lack of diversity in 3D RNA datasets, we explore training with structural clustering and cropping augmentations. Additionally, we define a suite of evaluation metrics to measure whether the generated RNA structures are globally self-consistent (via inverse folding followed by forward folding) and locally recover RNA-specific structural descriptors. The most performant version of RNA-FrameFlow generates locally realistic RNA backbones of 40-150 nucleotides, over 40% of which pass our validity criteria as measured by a self-consistency TM-score >= 0.45, at which two RNAs have the same global fold. Open-source code: https://github.com/rish-16/rna-backbone-design, Comment: To be presented as an Oral at ICML 2024 Structured Probabilistic Inference & Generative Modeling Workshop, and a Spotlight at ICML 2024 AI4Science Workshop

Consider a first-come, first-served single server queue with an initial workload $x>0$ and customers who arrive according to an inhomogeneous Poisson process with rate function $\lambda:[0,\infty)\rightarrow[0,\lambda_h ]$ for some $\lambda_h>0$. For each $i\in\mathbb{N}$, let $S_i$ (resp., $Y_i$) be the service (resp., patience) time of the $i$'th customer and assume that $(S_1,Y_1),(S_2,Y_2),\ldots$ is an iid sequence of bivariate random vectors with non-negative coordinates. A customer joins if and only if his patience time is not less than his prospective waiting time (i.e., the left-limit of the workload process at his arrival epoch). Let $\tau(x)$ be the first time when the system becomes empty and let $N^*_\lambda(\cdot)$ be the arrival process of those who join the queue. In the present work we suggest a novel coupling technique which is applied to derive stochastic upper bounds for the functionals: \begin{equation*} \int_0^{\tau(x)}g\circ W_x(t){\rm d}t\ \ \text{and}\ \ \int_0^{\tau(x)}g\circ W_x(t){\rm d}N^*_\lambda(t)\,, \end{equation*} where $W_x(\cdot)$ is the workload process in the queue and $g(\cdot)$ is any lower semi-continuous function. We also demonstrate how to utilise these bounds via some examples under the additional assumption that $\lambda(\cdot)$ is periodic.

Large language models (LLMs) have demonstrated the potential to mimic human social intelligence. However, most studies focus on simplistic and static self-report or performance-based tests, which limits the depth and validity of the analysis. In this paper, we developed a novel framework, InterIntent, to assess LLMs' social intelligence by mapping their ability to understand and manage intentions in a game setting. We focus on four dimensions of social intelligence: situational awareness, self-regulation, self-awareness, and theory of mind. Each dimension is linked to a specific game task: intention selection, intention following, intention summarization, and intention guessing. Our findings indicate that while LLMs exhibit high proficiency in selecting intentions, achieving an accuracy of 88%, their ability to infer the intentions of others is significantly weaker, trailing human performance by 20%. Additionally, game performance correlates with intention understanding, highlighting the importance of the four components towards success in this game. These findings underline the crucial role of intention understanding in evaluating LLMs' social intelligence and highlight the potential of using social deduction games as a complex testbed to enhance LLM evaluation. InterIntent contributes a structured approach to bridging the evaluation gap in social intelligence within multiplayer games.

Large Language Models trained on code corpora (code-LLMs) have demonstrated impressive performance in various coding assistance tasks. However, despite their increased size and training dataset, code-LLMs still have limitations such as suggesting codes with syntactic errors, variable misuse etc. Some studies argue that code-LLMs perform well on coding tasks because they use self-attention and hidden representations to encode relations among input tokens. However, previous works have not studied what code properties are not encoded by code-LLMs. In this paper, we conduct a fine-grained analysis of attention maps and hidden representations of code-LLMs. Our study indicates that code-LLMs only encode relations among specific subsets of input tokens. Specifically, by categorizing input tokens into syntactic tokens and identifiers, we found that models encode relations among syntactic tokens and among identifiers, but they fail to encode relations between syntactic tokens and identifiers. We also found that fine-tuned models encode these relations poorly compared to their pre-trained counterparts. Additionally, larger models with billions of parameters encode significantly less information about code than models with only a few hundred million parameters.

We study the $(a:b)$ Maker-Breaker subgraph game played on the edges of the complete graph $K_n$ on $n$ vertices, $n,a,b \in \mathbb{N}$ where the goal of Maker is to build a copy of a specific fixed subgraph $H$. In our work this is a spanning graph with minimum degree $k=k(n)$, a connected spanning subgraph or a Hamiltonian subgraph. In the $(a:b)$ game in each round Maker chooses $a$ unclaimed edges of $K_n$ and Breaker chooses $b$ unclaimed edges. Maker wins, if he succeeds to build a copy of the subgraph under consideration, otherwise Breaker wins. For the $k$-minimum-degree, we present a winning strategy for Maker leading to a bound that generalizes a bound of Gebauer and Szab{\'o} for the $(1:b)$ case. Moreover, we give an explicit strategy for Breaker for $b >(1+o(1)) \frac{an}{a+\ln(n)}$ in case of $a=o\left(\sqrt{\frac{n}{\ln(n)}}\right)$ and $k=o(\ln(n))$. Note that this bound is the same as the Maker bound presented by Hefetz et al. (2012) for the $(a:b)$ connectivity game, which implies that the asymptotic optimal bias for this game is $\frac{an}{a+\ln(n)}$. This resolves the open problem stated by these authors. We also study the $(a:b)$ Hamiltonicity game in which Maker's goal is to create a Hamiltonian subgraph. For the $(1:b)$ variant Krivelevich proved that $\left(1+o(1) \right)\frac{n}{\ln n}$ is the exact threshold bias. Controlling Breaker's vertex degree in the $(a:b)$ Maker-Breaker minimum degree game enables us to the asymptotic optimal generalized threshold bias for the $(a:b)$-game, both for $a=o\left(\sqrt{\frac{n}{\ln n}} \right)$ and $a=\Omega\left(\sqrt{\frac{n}{\ln n}} \right)$.

With the continuous advancement of large language models (LLMs), it is essential to create new benchmarks to effectively evaluate their expanding capabilities and identify areas for improvement. This work focuses on multi-image reasoning, an emerging capability in state-of-the-art LLMs. We introduce ReMI, a dataset designed to assess LLMs' ability to Reason with Multiple Images. This dataset encompasses a diverse range of tasks, spanning various reasoning domains such as math, physics, logic, code, table/chart understanding, and spatial and temporal reasoning. It also covers a broad spectrum of characteristics found in multi-image reasoning scenarios. We have benchmarked several cutting-edge LLMs using ReMI and found a substantial gap between their performance and human-level proficiency. This highlights the challenges in multi-image reasoning and the need for further research. Our analysis also reveals the strengths and weaknesses of different models, shedding light on the types of reasoning that are currently attainable and areas where future models require improvement. To foster further research in this area, we are releasing ReMI publicly: https://huggingface.co/datasets/mehrankazemi/ReMI.

Argument Mining (AM) involves identifying and extracting Argumentative Components (ACs) and their corresponding Argumentative Relations (ARs). Most of the prior works have broken down these tasks into multiple sub-tasks. Existing end-to-end setups primarily use the dependency parsing approach. This work introduces a generative paradigm-based end-to-end framework argTANL. argTANL frames the argumentative structures into label-augmented text, called Augmented Natural Language (ANL). This framework jointly extracts both ACs and ARs from a given argumentative text. Additionally, this study explores the impact of Argumentative and Discourse markers on enhancing the model's performance within the proposed framework. Two distinct frameworks, Marker-Enhanced argTANL (ME-argTANL) and argTANL with specialized Marker-Based Fine-Tuning, are proposed to achieve this. Extensive experiments are conducted on three standard AM benchmarks to demonstrate the superior performance of the ME-argTANL.

Training a deep neural network often involves stochastic optimization, meaning each run will produce a different model. The seed used to initialize random elements of the optimization procedure heavily influences the quality of a trained model, which may be obscure from many commonly reported summary statistics, like accuracy. However, random seed is often not included in hyper-parameter optimization, perhaps because the relationship between seed and model quality is hard to describe. This work attempts to describe the relationship between deep net models trained with different random seeds and the behavior of the expected model. We adopt robust hypothesis testing to propose a novel summary statistic for network similarity, referred to as the $\alpha$-trimming level. We use the $\alpha$-trimming level to show that the empirical cumulative distribution function of an ensemble model created from a collection of trained models with different random seeds approximates the average of these functions as the number of models in the collection grows large. This insight provides guidance for how many random seeds should be sampled to ensure that an ensemble of these trained models is a reliable representative. We also show that the $\alpha$-trimming level is more expressive than different performance metrics like validation accuracy, churn, or expected calibration error when taken alone and may help with random seed selection in a more principled fashion. We demonstrate the value of the proposed statistic in real experiments and illustrate the advantage of fine-tuning over random seed with an experiment in transfer learning.

Context: In a freeze drying (FD) process, an accurate observation and control of the process parameters at critical stages are at high importance. Particularly accurate and timely identification of the critical temperature (TC) at the end of primary drying phase would lead to heating energy cost reduction and product waste elimination by ending the process before the product's micro-collapse stage. Aim: Within this work, a combination of novel techniques optical technique called Intelligent Laser Speckle Imaging (ILSI) in association with product image texture analysis, coupled to an electrical impedance technique called through vial impedance spectroscopy (TVIS) has been applied onto the real-time product states observation process to identify the specific structural product surface/subsurface characteristics and hence the micro-collapse stage. Method: 2 cycles - one providing a profile for standard approach (non-collapsed) and another with a temperature ramp through TC to micro-collapse, TVIS provides an assessment of the onset of micro-collapse through the assessment of the acceleration in drying rate whereas the ILSI with pattern recognition detects the change in microstructure

The dual thinking framework considers fast, intuitive processing and slower, logical processing. The perception of dual thinking in vision requires images where inferences from intuitive and logical processing differ. We introduce an adversarial dataset to provide evidence for the dual thinking framework in human vision, which also aids in studying the qualitative behavior of deep learning models. Our study also addresses a major criticism of using classification models as computational models of human vision by using instance segmentation models that localize objects. The evidence underscores the importance of shape in identifying instances in human vision and shows that deep learning models lack an understanding of sub-structures, as indicated by errors related to the position and number of sub-components. Additionally, the similarity in errors made by models and intuitive human processing indicates that models only address intuitive thinking in human vision.

Distributed hybrid rendering (DHR) is a real-time rendering approach that incorporates cloud-based ray tracing with locally rasterized graphics for interactive thin client metaverse and game applications. With cloud assistance, DHR can generate high-fidelity ray-traced graphics contents remotely and deliver them to thin clients with low graphics capability, including standalone extended reality devices and mobile phones, while maintaining interactive frame rates for users under adverse network conditions. DHR can already achieve the effect of ray-traced hard shadows that form with the occlusion of direct illumination. We enhance the realism of these shadows by softening their edges with the direction of rays traced and approximating the occlusion of indirect illumination by reconstructing ray-traced ambient occlusion with a modified version of spatiotemporal variance-guided filtering. Our technique uses only 20-30% of the bandwidth of remote rendering and is also tolerant of delays of up to 200 ms with only slight distortion to the shadows along object edges.

We propose a novel method for fast and accurate training of physics-informed neural networks (PINNs) to find solutions to boundary value problems (BVPs) and initial boundary value problems (IBVPs). By combining the methods of training deep neural networks (DNNs) and Extreme Learning Machines (ELMs), we develop a model which has the expressivity of DNNs with the fine-tuning ability of ELMs. We showcase the superiority of our proposed method by solving several BVPs and IBVPs which include linear and non-linear ordinary differential equations (ODEs), partial differential equations (PDEs) and coupled PDEs. The examples we consider include a stiff coupled ODE system where traditional numerical methods fail, a 3+1D non-linear PDE, Kovasznay flow and Taylor-Green vortex solutions to incompressible Navier-Stokes equations and pure advection solution of 1+1 D compressible Euler equation. The Theory of Functional Connections (TFC) is used to exactly impose initial and boundary conditions (IBCs) of (I)BVPs on PINNs. We propose a modification to the TFC framework named Reduced TFC and show a significant improvement in the training and inference time of PINNs compared to IBCs imposed using TFC. Furthermore, Reduced TFC is shown to be able to generalize to more complex boundary geometries which is not possible with TFC. We also introduce a method of applying boundary conditions at infinity for BVPs and numerically solve the pure advection in 1+1 D Euler equations using these boundary conditions., Comment: Accepted for publication in CNSNS

Artificial neural networks (ANNs) are powerful tools capable of approximating any arbitrary mathematical function, but their interpretability remains limited, rendering them as black box models. To address this issue, numerous methods have been proposed to enhance the explainability and interpretability of ANNs. In this study, we introduce the application of information geometric framework to investigate phase transition-like behavior during the training of ANNs and relate these transitions to overfitting in certain models. The evolution of ANNs during training is studied by looking at the probability distribution of its parameters. Information geometry utilizing the principles of differential geometry, offers a unique perspective on probability and statistics by considering probability density functions as points on a Riemannian manifold. We create this manifold using a metric based on Fisher information to define a distance and a velocity. By parameterizing this distance and velocity with training steps, we study how the ANN evolves as training progresses. Utilizing standard datasets like MNIST, FMNIST and CIFAR-10, we observe a transition in the motion on the manifold while training the ANN and this transition is identified with over-fitting in the ANN models considered. The information geometric transitions observed is shown to be mathematically similar to the phase transitions in physics. Preliminary results showing finite-size scaling behavior is also provided. This work contributes to the development of robust tools for improving the explainability and interpretability of ANNs, aiding in our understanding of the variability of the parameters these complex models exhibit during training., Comment: Accepted for publication in Neurocomputing journal

There is vivid research on adapting Large Language Models (LLMs) to perform a variety of tasks in high-stakes domains such as healthcare. Despite their popularity, there is a lack of understanding of the extent and contributing factors that allow LLMs to recall relevant knowledge and combine it with presented information in the clinical and biomedical domain: a fundamental pre-requisite for success on down-stream tasks. Addressing this gap, we use Multiple Choice and Abstractive Question Answering to conduct a large-scale empirical study on 22 datasets in three generalist and three specialist biomedical sub-domains. Our multifaceted analysis of the performance of 15 LLMs, further broken down by sub-domain, source of knowledge and model architecture, uncovers success factors such as instruction tuning that lead to improved recall and comprehension. We further show that while recently proposed domain-adapted models may lack adequate knowledge, directly fine-tuning on our collected medical knowledge datasets shows encouraging results, even generalising to unseen specialist sub-domains. We complement the quantitative results with a skill-oriented manual error analysis, which reveals a significant gap between the models' capabilities to simply recall necessary knowledge and to integrate it with the presented context. To foster research and collaboration in this field we share M-QALM, our resources, standardised methodology, and evaluation results, with the research community to facilitate further advancements in clinical knowledge representation learning within language models., Comment: Accepted at ACL 2024 (Findings)

LiteBIRD, the next-generation cosmic microwave background (CMB) experiment, aims for a launch in Japan's fiscal year 2032, marking a major advancement in the exploration of primordial cosmology and fundamental physics. Orbiting the Sun-Earth Lagrangian point L2, this JAXA-led strategic L-class mission will conduct a comprehensive mapping of the CMB polarization across the entire sky. During its 3-year mission, LiteBIRD will employ three telescopes within 15 unique frequency bands (ranging from 34 through 448 GHz), targeting a sensitivity of 2.2\,$\mu$K-arcmin and a resolution of 0.5$^\circ$ at 100\,GHz. Its primary goal is to measure the tensor-to-scalar ratio $r$ with an uncertainty $\delta r = 0.001$, including systematic errors and margin. If $r \geq 0.01$, LiteBIRD expects to achieve a $>5\sigma$ detection in the $\ell=$2-10 and $\ell=$11-200 ranges separately, providing crucial insight into the early Universe. We describe LiteBIRD's scientific objectives, the application of systems engineering to mission requirements, the anticipated scientific impact, and the operations and scanning strategies vital to minimizing systematic effects. We will also highlight LiteBIRD's synergies with concurrent CMB projects., Comment: 23 pages, 9 figures, 1 table, SPIE Astronomical Telescopes + Instrumentation 2024

$ $Let $k$ be a field of characteristic zero. If $c_1, c_2\in k\setminus \{0\}, s,t\geq 1$ and $u\geq 0$, then it is shown that the $k$-derivations $\partial_x + x^u(c_1x^ty^s+c_2)\partial_y$ and $\partial_x + x^u(c_1x^t+c_2y^{s+1})\partial_y$ of $k[x,y]$ are simple. We also give a necessary and sufficient condition for the $k$-derivation $y^r\partial_x + (c_1x^{t_1}y^{s_1}+c_2x^{t_2}y^{s_2})\partial_y$, where $r, t_1, s_1, t_2, s_2 \geq 0$ and $c_1, c_2\in k$, of $k[x,y]$ to be simple.

Classical machine learning, extensively utilized across diverse domains, faces limitations in speed, efficiency, parallelism, and processing of complex datasets. In contrast, quantum machine learning algorithms offer significant advantages, including exponentially faster computations, enhanced data handling capabilities, inherent parallelism, and improved optimization for complex problems. In this study, we used the entanglement-enhanced quantum kernel in quantum support vector machine to train complex respiratory data sets. Compared to classical algorithms, our findings reveal that QSVM performs better with 45% higher accuracy for complex respiratory data sets while maintaining comparable performance with linear datasets in contrast to their classical counterparts executed on a 2-qubit system. Through our study, we investigate the efficacy of the QSVM-Kernel algorithm in harnessing the enhanced dimensionality of the quantum Hilbert space for effectively training complex datasets.

Hierarchical model-based reinforcement learning (HMBRL) aims to combine the benefits of better sample efficiency of model based reinforcement learning (MBRL) with the abstraction capability of hierarchical reinforcement learning (HRL) to solve complex tasks efficiently. While HMBRL has great potential, it still lacks wide adoption. In this work we describe a novel HMBRL framework and evaluate it thoroughly. To complement the multi-layered decision making idiom characteristic for HRL, we construct hierarchical world models that simulate environment dynamics at various levels of temporal abstraction. These models are used to train a stack of agents that communicate in a top-down manner by proposing goals to their subordinate agents. A significant focus of this study is the exploration of a static and environment agnostic temporal abstraction, which allows concurrent training of models and agents throughout the hierarchy. Unlike most goal-conditioned H(MB)RL approaches, it also leads to comparatively low dimensional abstract actions. Although our HMBRL approach did not outperform traditional methods in terms of final episode returns, it successfully facilitated decision making across two levels of abstraction using compact, low dimensional abstract actions. A central challenge in enhancing our method's performance, as uncovered through comprehensive experimentation, is model exploitation on the abstract level of our world model stack. We provide an in depth examination of this issue, discussing its implications for the field and suggesting directions for future research to overcome this challenge. By sharing these findings, we aim to contribute to the broader discourse on refining HMBRL methodologies and to assist in the development of more effective autonomous learning systems for complex decision-making environments., Comment: 26 pages, 14 figures

Image relighting is the task of showing what a scene from a source image would look like if illuminated differently. Inverse graphics schemes recover an explicit representation of geometry and a set of chosen intrinsics, then relight with some form of renderer. However error control for inverse graphics is difficult, and inverse graphics methods can represent only the effects of the chosen intrinsics. This paper describes a relighting method that is entirely data-driven, where intrinsics and lighting are each represented as latent variables. Our approach produces SOTA relightings of real scenes, as measured by standard metrics. We show that albedo can be recovered from our latent intrinsics without using any example albedos, and that the albedos recovered are competitive with SOTA methods.

Fitch Cheney's 5-card trick was introduced in 1950. In 2013, Mulcahy invented a 4-card trick in which the cards are allowed to be displayed face down. We suggest our own invention: a 3-card trick in which the cards can be face down and also allowed to be placed both vertically and horizontally. We discuss the theory behind all the tricks and estimate the maximum deck size given the number of chosen cards. We also discuss the cases of hiding several cards and the deck that has duplicates., Comment: 24 pages, 13 tables

Language models have the ability to perform in-context learning (ICL), allowing them to flexibly adapt their behavior based on context. This contrasts with in-weights learning, where information is statically encoded in model parameters from iterated observations of the data. Despite this apparent ability to learn in-context, language models are known to struggle when faced with unseen or rarely seen tokens. Hence, we study $\textbf{structural in-context learning}$, which we define as the ability of a model to execute in-context learning on arbitrary tokens -- so called because the model must generalize on the basis of e.g. sentence structure or task structure, rather than semantic content encoded in token embeddings. An ideal model would be able to do both: flexibly deploy in-weights operations (in order to robustly accommodate ambiguous or unknown contexts using encoded semantic information) and structural in-context operations (in order to accommodate novel tokens). We study structural in-context algorithms in a simple part-of-speech setting using both practical and toy models. We find that active forgetting, a technique that was recently introduced to help models generalize to new languages, forces models to adopt structural in-context learning solutions. Finally, we introduce $\textbf{temporary forgetting}$, a straightforward extension of active forgetting that enables one to control how much a model relies on in-weights vs. in-context solutions. Importantly, temporary forgetting allows us to induce a $\textit{dual process strategy}$ where in-context and in-weights solutions coexist within a single model., Comment: 9 pages, 5 figures

Numerous algorithms have been proposed to $\textit{align}$ language models to remove undesirable behaviors. However, the challenges associated with a very large state space and creating a proper reward function often result in various jailbreaks. Our paper aims to examine this effect of reward in the controlled setting of positive sentiment language generation. Instead of online training of a reward model based on human feedback, we employ a statically learned sentiment classifier. We also consider a setting where our model's weights and activations are exposed to an end-user after training. We examine a pretrained GPT-2 through the lens of mechanistic interpretability before and after proximal policy optimization (PPO) has been applied to promote positive sentiment responses. Using these insights, we (1) attempt to "hack" the PPO-ed model to generate negative sentiment responses and (2) add a term to the reward function to try and alter `negative' weights., Comment: 8 pages, 4 figures

In this paper, we explore the interplay between Predictive Control and closed-loop optimality, spanning from Model Predictive Control to Data-Driven Predictive Control. Predictive Control in general relies on some form of prediction scheme on the real system trajectories. However, these predictions may not accurately capture the real system dynamics, for e.g., due to stochasticity, resulting in sub-optimal control policies. This lack of optimality is a critical issue in case of problems with economic objectives. We address this by providing sufficient conditions on the underlying prediction scheme such that a Predictive Controller can achieve closed-loop optimality. However, these conditions do not readily extend to Data-Driven Predictive Control. In this context of closed-loop optimality, we conclude that the factor distinguishing the approaches within Data-Driven Predictive Control is if they can be cast as a sequential decision-making process or not, rather than the dichotomy of model-based vs. model-free. Furthermore, we show that the conventional approach of improving the prediction accuracy from data may not guarantee optimality.

Current state-of-the-art synchrony-based models encode object bindings with complex-valued activations and compute with real-valued weights in feedforward architectures. We argue for the computational advantages of a recurrent architecture with complex-valued weights. We propose a fully convolutional autoencoder, SynCx, that performs iterative constraint satisfaction: at each iteration, a hidden layer bottleneck encodes statistically regular configurations of features in particular phase relationships; over iterations, local constraints propagate and the model converges to a globally consistent configuration of phase assignments. Binding is achieved simply by the matrix-vector product operation between complex-valued weights and activations, without the need for additional mechanisms that have been incorporated into current synchrony-based models. SynCx outperforms or is strongly competitive with current models for unsupervised object discovery. SynCx also avoids certain systematic grouping errors of current models, such as the inability to separate similarly colored objects without additional supervision., Comment: NeurIPS 2024 camera-ready

Existing AI alignment approaches assume that preferences are static, which is unrealistic: our preferences change, and may even be influenced by our interactions with AI systems themselves. To clarify the consequences of incorrectly assuming static preferences, we introduce Dynamic Reward Markov Decision Processes (DR-MDPs), which explicitly model preference changes and the AI's influence on them. We show that despite its convenience, the static-preference assumption may undermine the soundness of existing alignment techniques, leading them to implicitly reward AI systems for influencing user preferences in ways users may not truly want. We then explore potential solutions. First, we offer a unifying perspective on how an agent's optimization horizon may partially help reduce undesirable AI influence. Then, we formalize different notions of AI alignment that account for preference change from the outset. Comparing the strengths and limitations of 8 such notions of alignment, we find that they all either err towards causing undesirable AI influence, or are overly risk-averse, suggesting that a straightforward solution to the problems of changing preferences may not exist. As there is no avoiding grappling with changing preferences in real-world settings, this makes it all the more important to handle these issues with care, balancing risks and capabilities. We hope our work can provide conceptual clarity and constitute a first step towards AI alignment practices which explicitly account for (and contend with) the changing and influenceable nature of human preferences., Comment: Accepted to ICML 2024

We present the technical details on how large-scale structure (LSS) catalogs are constructed from redshifts measured from spectra observed by the Dark Energy Spectroscopic Instrument (DESI). The LSS catalogs provide the information needed to determine the relative number density of DESI tracers as a function of redshift and celestial coordinates and, e.g., determine clustering statistics. We produce catalogs that are weighted subsamples of the observed data, each matched to a weighted `random' catalog that forms an unclustered sampling of the probability density that DESI could have observed those data at each location. Precise knowledge of the DESI observing history and associated hardware performance allows for a determination of the DESI footprint and the number of times DESI has covered it at sub-arcsecond level precision. This enables the completeness of any DESI sample to be modeled at this same resolution. The pipeline developed to create LSS catalogs has been designed to easily allow robustness tests and enable future improvements. We describe how it allows ongoing work improving the match between galaxy and random catalogs, such as including further information when assigning redshifts to randoms, accounting for fluctuations in target density, accounting for variation in the redshift success rate, and accommodating blinding schemes., Comment: Accepted (by JCAP) version of supporting publication of DESI 2024II: Sample definitions, characteristics, and two-point clustering statistics

A central challenge in quantum information science and technology is achieving real-time estimation and feedforward control of quantum systems. This challenge is compounded by the inherent inhomogeneity of quantum resources, such as qubit properties and controls, and their intrinsically probabilistic nature. This leads to stochastic challenges in error detection and probabilistic outcomes in processes such as heralded remote entanglement. Given these complexities, optimizing the construction of quantum resource states is an NP-hard problem. In this paper, we address the quantum resource scheduling issue by formulating the problem and simulating it within a digitized environment, allowing the exploration and development of agent-based optimization strategies. We employ reinforcement learning agents within this probabilistic setting and introduce a new framework utilizing a Transformer model that emphasizes self-attention mechanisms for pairs of qubits. This approach facilitates dynamic scheduling by providing real-time, next-step guidance. Our method significantly improves the performance of quantum systems, achieving more than a 3$\times$ improvement over rule-based agents, and establishes an innovative framework that improves the joint design of physical and control systems for quantum applications in communication, networking, and computing.

Experimental control involves a lot of manual effort with non-trivial decisions for precise adjustments. Here, we study the automatic experimental alignment for coupling laser light into an optical fiber using reinforcement learning (RL). We face several real-world challenges, such as time-consuming training, partial observability, and noisy actions due to imprecision in the mirror steering motors. We show that we can overcome these challenges: To save time, we use a virtual testbed to tune our environment for dealing with partial observability and use relatively sample-efficient model-free RL algorithms like Soft Actor-Critic (SAC) or Truncated Quantile Critics (TQC). Furthermore, by fully training on the experiment, the agent learns directly to handle the noise present. In our extensive experimentation, we show that we are able to achieve 90% coupling, showcasing the effectiveness of our proposed approaches. We reach this efficiency, which is comparable to that of a human expert, without additional feedback loops despite the motors' inaccuracies. Our result is an example of the readiness of RL for real-world tasks. We consider RL a promising tool for reducing the workload in labs., Comment: 10 pages + 10 pages appendices, 3 + 11 figures

Machine learning (ML) has recently shown significant promise in modelling atmospheric systems, such as the weather. Many of these ML models are autoregressive, and error accumulation in their forecasts is a key problem. However, there is no clear definition of what `error accumulation' actually entails. In this paper, we propose a definition and an associated metric to measure it. Our definition distinguishes between errors which are due to model deficiencies, which we may hope to fix, and those due to the intrinsic properties of atmospheric systems (chaos, unobserved variables), which are not fixable. We illustrate the usefulness of this definition by proposing a simple regularization loss penalty inspired by it. This approach shows performance improvements (according to RMSE and spread/skill) in a selection of atmospheric systems, including the real-world weather prediction task., Comment: Submitted to NeurIPS 2024. 27 pages (including appendices)

The advancement of large language models has significantly improved natural language processing. However, challenges such as jailbreaks (prompt injections that cause an LLM to follow instructions contrary to its intended use), hallucinations (generating incorrect or misleading information), and comprehension errors remain prevalent. In this report, we present a comparative analysis of the performance of fifteen distinct models, with each model undergoing a standardized test comprising 38 queries across three key metrics: jailbreaks, hallucinations, and comprehension errors. The models are assessed based on the total occurrences of jailbreaks, hallucinations, and comprehension errors. Our work exposes these models' inherent vulnerabilities and challenges the notion of human-level language comprehension of these models. We have empirically analysed the impact of non-standard Unicode characters on LLMs and their safeguarding mechanisms on the best-performing LLMs, including GPT-4, Gemini 1.5 Pro, LlaMA-3-70B, and Claude 3 Opus. By incorporating alphanumeric symbols from Unicode outside the standard Latin block and variants of characters in other languages, we observed a reduction in the efficacy of guardrails implemented through Reinforcement Learning Human Feedback (RLHF). Consequently, these models exhibit heightened vulnerability to content policy breaches and prompt leakage. Our study also suggests a need to incorporate non-standard Unicode text in LLM training data to enhance the capabilities of these models., Comment: 46 pages

A major challenge facing the world is the provision of equitable and universal access to quality education. Recent advances in generative AI (gen AI) have created excitement about the potential of new technologies to offer a personal tutor for every learner and a teaching assistant for every teacher. The full extent of this dream, however, has not yet materialised. We argue that this is primarily due to the difficulties with verbalising pedagogical intuitions into gen AI prompts and the lack of good evaluation practices, reinforced by the challenges in defining excellent pedagogy. Here we present our work collaborating with learners and educators to translate high level principles from learning science into a pragmatic set of seven diverse educational benchmarks, spanning quantitative, qualitative, automatic and human evaluations; and to develop a new set of fine-tuning datasets to improve the pedagogical capabilities of Gemini, introducing LearnLM-Tutor. Our evaluations show that LearnLM-Tutor is consistently preferred over a prompt tuned Gemini by educators and learners on a number of pedagogical dimensions. We hope that this work can serve as a first step towards developing a comprehensive educational evaluation framework, and that this can enable rapid progress within the AI and EdTech communities towards maximising the positive impact of gen AI in education.

We construct a Lagrangian multiform for the class of cyclotomic (rational) Gaudin models by formulating its hierarchy within the Lie dialgebra framework of Semenov-Tian-Shansky and by using the framework of Lagrangian multiforms on coadjoint orbits. This provides the first example of a Lagrangian multiform for an integrable hierarchy whose classical $r$-matrix is non-skew-symmetric and spectral parameter-dependent. As an important by-product of the construction, we obtain a Lagrangian multiform for the periodic Toda chain by choosing an appropriate realisation of the cyclotomic Gaudin Lax matrix. This fills a gap in the landscape of Toda models as only the open and infinite chains had been previously cast into the Lagrangian multiform framework. A slightly different choice of realisation produces the so-called discrete self-trapping (DST) model. We demonstrate the versatility of the framework by coupling the periodic Toda chain with the DST model and by obtaining a Lagrangian multiform for the corresponding integrable hierarchy.

We adapt the notion of a (relatively) definable subset of Aut(M) when M is a saturated model to the case Aut(M/A) when M is atomic and strongly omega-homogeneous over A. We discuss the existence and uniqueness of invariant measures on the Boolean algebra of definable subsets of Aut(M/A). For example when Th(M) is stable we have existence and uniqueness. We also discuss the compatibility of our definability notions with definable Galois cohomology and differential Galois theory., Comment: 16 pages

During the first half of the fourth observing run (O4a) of the International Gravitational Wave Network (IGWN), the Zwicky Transient Facility (ZTF) conducted a systematic search for kilonova (KN) counterparts to binary neutron star (BNS) and neutron star-black hole (NSBH) merger candidates. Here, we present a comprehensive study of the five high-significance (FAR < 1 per year) BNS and NSBH candidates in O4a. Our follow-up campaigns relied on both target-of-opportunity observations (ToO) and re-weighting of the nominal survey schedule to maximize coverage. We describe the toolkit we have been developing, Fritz, an instance of SkyPortal, instrumental in coordinating and managing our telescope scheduling, candidate vetting, and follow-up observations through a user-friendly interface. ZTF covered a total of 2841 deg$^2$ within the skymaps of the high-significance GW events, reaching a median depth of g~20.2 mag. We circulated 15 candidates, but found no viable KN counterpart to any of the GW events. Based on the ZTF non-detections of the high-significance events in O4a, we used a Bayesian approach, nimbus, to quantify the posterior probability of KN model parameters that are consistent with our non-detections. Our analysis favors KNe with initial absolute magnitude fainter than -16 mag. The joint posterior probability of a GW170817-like KN associated with all our O4a follow-ups was 64%. Additionally, we use a survey simulation software, simsurvey, to determine that our combined filtered efficiency to detect a GW170817-like KN is 36%, when considering the 5 confirmed astrophysical events in O3 (1 BNS and 4 NSBH), along with our O4a follow-ups. Following Kasliwal et al. (2020), we derived joint constraints on the underlying KN luminosity function based on our O3 and O4a follow-ups, determining that no more than 76% of KNe fading at 1 mag/day can peak at a magnitude brighter than -17.5 mag., Comment: submitted

A K-mirror is a device that rotates the wavefront of an incident optical field. It has recently gained prominence over Dove prism, another commonly used wavefront rotator, due to the fact that while a K-mirror has several controls for adjusting the internal reflections, a Dove prism is made of a single glass element with no additional control. Thus, one can obtain much lower angular deviations of transmitting wavefronts using a K-mirror than with a Dove prism. However, the accompanying polarization changes in the transmitted field due to rotation persist even in the commercially available K-mirrors. A recent theoretical work [Applied Optics, 61, 8302 (2022)] shows that it is possible to optimize the base angle of a K-mirror for a given refractive index such that the accompanying polarization changes are minimum. In contrast, we show in this article that by optimizing the refractive index it is possible to design a K-mirror at any given base angle and with any given value for the mean polarization change, including near-zero values. Furthermore, we experimentally demonstrate a K-mirror with an order-of-magnitude lower mean polarization change than that of the commercially available K-mirrors. This can have important practical implications for OAM-based applications that require precise wavefront rotation control., Comment: Manuscript: 9 pages, 9 figures

We compute equivariant fundamental classes of orbits in GL(2)-representations. As applications, we find degrees of the orbit closures corresponding to elliptic fibrations and self-maps of the projective line., Comment: 1 table, 3 figures. Comments welcome!

Well-structured transition systems (WSTS) are an abstract family of systems that encompasses a vast landscape of infinite-state systems. By requiring a well-quasi-ordering (wqo) on the set of states, a WSTS enables generic algorithms for classic verification tasks such as coverability and termination. However, even for systems that are WSTS like vector addition systems (VAS), the framework is notoriously ill-equipped to analyse reachability (as opposed to coverability). Moreover, some important types of infinite-state systems fall out of WSTS' scope entirely, such as pushdown systems (PDS). Inspired by recent algorithmic techniques on VAS, we propose an abstract notion of systems where the set of runs is equipped with a wqo and supports amalgamation of runs. We show that it subsumes a large class of infinite-state systems, including (reachability languages of) VAS and PDS, and even all systems from the abstract framework of valence systems, except for those already known to be Turing-complete. Moreover, this abstract setting enables simple and general algorithmic solutions to unboundedness problems, which have received much attention in recent years. We present algorithms for the (i) simultaneous unboundedness problem (which implies computability of downward closures and decidability of separability by piecewise testable languages), (ii) computing priority downward closures, (iii) deciding whether a language is bounded, meaning included in $w_1^*\cdots w_k^*$ for some words $w_1,\ldots,w_k$, and (iv) effective regularity of unary languages. This leads to either drastically simpler proofs or new decidability results for a rich variety of systems., Comment: Erratum: Updated test for negative SUP instances in Section 4.1

We explore the influence of radio-mode feedback on the properties of the cool circumgalactic medium (CGM). To this end, we assemble a statistical sample of approximately 30,000 radio galaxies with background quasars by combining optical spectroscopic measurements of luminous red galaxies (LRGs) and quasars from the year 1 dataset of Dark Energy Spectroscopic Instrument (DESI) and radio sources from the LOw-Frequency ARray Two-metre Sky Survey (LoTSS) DR2 catalog and the Very Large Array Sky Survey (VLASS) quick look catalog. Galaxies with similar optical properties but with no radio counterparts in LoTSS and VLASS are selected as the control group. We measure the cool CGM properties of radio galaxies and their control samples traced by MgII absorption lines, including covering fraction, rest equivalent width, and gas kinematics. Our results show no significant difference in the properties of gas around radio galaxies and their control sample, indicating that the operating radio-mode feedback of massive galaxies does not produce detectable effects on the properties of the cool CGM. Finally, we show that the CGM of radio galaxies contain a non-negligible amount of cool gas with approximately 10^10 solar masses. This abundance can place a stringent constraint on the radio-mode feedback models., Comment: 20 pages, 12 figures

Neural ranking models have become increasingly popular for real-world search and recommendation systems in recent years. Unlike their tree-based counterparts, neural models are much less interpretable. That is, it is very difficult to understand their inner workings and answer questions like how do they make their ranking decisions? or what document features do they find important? This is particularly disadvantageous since interpretability is highly important for real-world systems. In this work, we explore feature selection for neural learning-to-rank (LTR). In particular, we investigate six widely-used methods from the field of interpretable machine learning (ML) and introduce our own modification, to select the input features that are most important to the ranking behavior. To understand whether these methods are useful for practitioners, we further study whether they contribute to efficiency enhancement. Our experimental results reveal a large feature redundancy in several LTR benchmarks: the local selection method TabNet can achieve optimal ranking performance with less than 10 features; the global methods, particularly our G-L2X, require slightly more selected features, but exhibit higher potential in improving efficiency. We hope that our analysis of these feature selection methods will bring the fields of interpretable ML and LTR closer together., Comment: Published at ECIR 2024 as a long paper. 13 pages excl. reference, 20 pages incl. reference

Future real-time applications like smart cities will use complex Machine Learning (ML) models for a variety of tasks. Timely status information is required for these applications to be reliable. Offloading computation to a mobile edge cloud (MEC) can reduce the completion time of these tasks. However, using the MEC may come at a cost such as related to use of a cloud service or privacy. In this paper, we consider a source that generates time-stamped status updates for delivery to a monitor after processing by the mobile device or MEC. We study how a scheduler must forward these updates to achieve timely updates at the monitor but also limit MEC usage. We measure timeliness at the monitor using the age of information (AoI) metric. We formulate this problem as an infinite horizon Markov decision process (MDP) with an average cost criterion. We prove that an optimal scheduling policy has an age-threshold structure that depends on how long an update has been in service.

Machine learning interatomic potentials (MLIPs) have introduced a new paradigm for atomic simulations. Recent advancements have seen the emergence of universal MLIPs (uMLIPs) that are pre-trained on diverse materials datasets, providing opportunities for both ready-to-use universal force fields and robust foundations for downstream machine learning refinements. However, their performance in extrapolating to out-of-distribution complex atomic environments remains unclear. In this study, we highlight a consistent potential energy surface (PES) softening effect in three uMLIPs: M3GNet, CHGNet, and MACE-MP-0, which is characterized by energy and force under-prediction in a series of atomic-modeling benchmarks including surfaces, defects, solid-solution energetics, phonon vibration modes, ion migration barriers, and general high-energy states. We find that the PES softening behavior originates from a systematic underprediction error of the PES curvature, which derives from the biased sampling of near-equilibrium atomic arrangements in uMLIP pre-training datasets. We demonstrate that the PES softening issue can be effectively rectified by fine-tuning with a single additional data point. Our findings suggest that a considerable fraction of uMLIP errors are highly systematic, and can therefore be efficiently corrected. This result rationalizes the data-efficient fine-tuning performance boost commonly observed with foundational MLIPs. We argue for the importance of a comprehensive materials dataset with improved PES sampling for next-generation foundational MLIPs.

In this paper, we introduce saturation and semisaturation functions of sequences, and we prove a number of fundamental results about these functions. Given a forbidden sequence $u$ with $r$ distinct letters, we say that a sequence $s$ on a given alphabet is $u$-saturated if $s$ is $r$-sparse, $u$-free, and adding any letter from the alphabet to an arbitrary position in $s$ violates $r$-sparsity or induces a copy of $u$. We say that $s$ is $u$-semisaturated if $s$ is $r$-sparse and adding any letter from the alphabet to $s$ violates $r$-sparsity or induces a new copy of $u$. Let the saturation function $\operatorname{Sat}(u, n)$ denote the minimum possible length of a $u$-saturated sequence on an alphabet of size $n$, and let the semisaturation function $\operatorname{Ssat}(u, n)$ denote the minimum possible length of a $u$-semisaturated sequence on an alphabet of size $n$. For alternating sequences, we determine both the saturation function and the semisaturation function up to a constant multiplicative factor. We show for every sequence that the semisaturation function is always either $O(1)$ or $\Theta(n)$. For the saturation function, we show that every sequence $u$ has either $\operatorname{Sat}(u, n) \ge n$ or $\operatorname{Sat}(u, n) = O(1)$. For every sequence with $2$ distinct letters, we show that the saturation function is always either $O(1)$ or $\Theta(n)$.

Hexagonal boron nitride (h-BN) is brittle, however, its atomic-scale structural engineering can lead to unprecedented physical properties. Here we report the bulk synthesis of high-density crystalline h-BN solids by using high-temperature spark plasma sintering (SPS) of micron size h-BN powders. In addition to the high mechanical strength and ductile response of such materials, we have obtained anomalous values of dielectric constant beyond theoretical limits, high thermal conductivity, and exceptional neutron radiation shielding capability. Through exhaustive characterizations we reveal that SPS induces non-basal plane crystallinity, twisting of layers, and facilitates inter-grain fusion with a high degree of in-plane alignment across macroscale dimensions, resulting in near-theoretical density and anomalous properties. Our findings highlight the importance of material design, via new approaches such as twisting and interconnections between atomically thin layers, to create novel ceramics with properties that could go beyond their intrinsic theoretical predictions., Comment: Authors revised version, 46 pages, 4 figures

We apply the 3D dust radiative transfer code SKIRT to the low-redshift ($z\leq0.1$) galaxy population in the TNG100 cosmological simulation, the fiducial run of the IllustrisTNG project. We compute global fluxes and spectral energy distributions (SEDs) from the far-ultraviolet to the sub-millimeter for $\approx\,$60 000 galaxies. Our post-processing methodology follows the study of Tr\v{c}ka et al. (2022) of the higher-resolution TNG50 simulation. We verify that TNG100 reproduces observational luminosity functions at low redshifts to excellent precision, unlike TNG50. Additionally, we test the realism of our TNG100 plus SKIRT fluxes by comparing various flux and color relations to data from the GAMA survey. TNG100 broadly reproduces the observed distributions, but we predict ultraviolet colors that are too blue by $\approx\,$0.4 mag, possibly related to the extinction in the star-forming regions subgrid model not being selective enough. Furthermore, we find that the simulated galaxies exhibit mid-infrared fluxes elevated by up to $\approx\,$0.5 mag that we attribute to overly effective stochastic heating of the diffuse dust. All synthetic broadband fluxes and SEDs are made publicly available in three orientations and four apertures, and can readily be used to study TNG100 galaxies in a mock observational fashion., Comment: Main text 14 pages, 5 figures, accepted to MNRAS. Generated data products are publicly available at www.tng-project.org/gebek24

Maker-Breaker subgraph games are among the most famous combinatorial games. For given $n,q \in \mathbb{N}$ and a subgraph $C$ of the complete graph $K_n$, the two players, called Maker and Breaker, alternately claim edges of $K_n$. In each round of the game Maker claims one edge and Breaker is allowed to claim up to $q$ edges. If Maker is able to claim all edges of a copy of $C$, he wins the game. Otherwise Breaker wins. In this work we introduce the first constructive strategy for Maker for the $C_4$-Maker-Breaker game and show that he can win the game if $q < 0.16 n^{2/3}$. According to the theorem of Bednarska and Luczak (2000) $n^{2/3}$ is asymptotically optimal for this game, but the constant given there for a random Maker strategy is magnitudes apart from our constant 0.16.

Differences between the local value of the Hubble constant measured via the distance ladder versus the value inferred from the cosmic microwave background with the assumption of the standard $\Lambda$CDM model have reached over 5$\sigma$ significance. To determine if this discrepancy is due to new physics or more mundane systematic errors, it is essential to remove as many sources of systematic uncertainty as possible by developing high-precision distance ladders that are independent of the traditional Cepheid and Type Ia supernovae route. Here we present JWST observations of three early-type Fornax Cluster galaxies, the first of fourteen observations from a Cycle 2 JWST program. Our modest integration times allow us to measure highly precise tip of the red giant branch (TRGB) distances, and will also be used to perform measurements of Surface Brightness Fluctuations (SBF). From these three galaxies, we determine an average TRGB distance modulus to the Fornax Cluster of $\mu$ = 31.424 $\pm$ 0.077 mag, or D = 19.3 $\pm$ 0.7 Mpc. With eleven more scheduled observations in nearby elliptical galaxies, our program will allow us set the zero point of the SBF scale to better than 2$\%$ for more distant measurements, charting a path towards a high-precision measurement of $H_{0}$ that is independent of the traditional Cepheid-SN Ia distance ladder., Comment: 14 pages, 9 figures, accepted to ApJ

The baryon mass content of dark matter halos in the early Universe depends on global factors - e.g. ionising ultraviolet (UV) radiation background - and local factors - e.g. star formation efficiency and assembly history. We use a lightweight semi-analytical model to investigate how local and global factors impact halo baryon mass content at redshifts of $z\geq 5$. Our model incorporates a time delay between when stars form and when they produce feedback, which drive bursts of star formation, and a mass and redshift dependent UV background, which captures the influence of cosmological reionization on gas accretion onto halos. We use statistically representative halo assembly histories and assume that the cosmological gas accretion rate is proportional to the halo mass accretion rate. Delayed feedback leads to oscillations in gas mass with cosmic time, behaviour that cannot be captured with instantaneous feedback. Highly efficient star formation drives stronger oscillations, while strong feedback impacts when oscillations occur; in contrast, inefficient star formation and weak feedback produce similar long-term behaviour to that observed in instantaneous feedback models. If the delayed feedback timescale is too long, a halo retains its gas reservoir but the feedback suppresses star formation. Our model predicts that lower mass systems ($\leq 10^7 \text{M}_\odot$) at $z \leq 10$ should be strongly gas deficient, whereas higher mass systems retain their gas reservoirs because they are sufficiently massive to continue accreting gas through cosmological reionization. Interestingly, in higher mass halos, the median $m_\star/(m_\star+m_\text{g}) \simeq 0.01-0.05$, but is a factor of 3-5 smaller when feedback is delayed. Our model does not include seed supermassive black hole feedback, which is necessary to explain massive quenched galaxies in the early Universe., Comment: 12 pages, 9 figures. Accepted for publication in Publications of the Astronomical Society of Australia

When multitudes of features can plausibly be associated with a response, both privacy considerations and model parsimony suggest grouping them to increase the predictive power of a regression model. Specifically, the identification of groups of predictors significantly associated with the response variable eases further downstream analysis and decision-making. This paper proposes a new data analysis methodology that utilizes the high-dimensional predictor space to construct an implicit network with weighted edges %and weights on the edges to identify significant associations between the response and the predictors. Using a population model for groups of predictors defined via network-wide metrics, a new supervised grouping algorithm is proposed to determine the correct group, with probability tending to one as the sample size diverges to infinity. For this reason, we establish several theoretical properties of the estimates of network-wide metrics. A novel model-assisted bootstrap procedure that substantially decreases computational complexity is developed, facilitating the assessment of uncertainty in the estimates of network-wide metrics. The proposed methods account for several challenges that arise in the high-dimensional data setting, including (i) a large number of predictors, (ii) uncertainty regarding the true statistical model, and (iii) model selection variability. The performance of the proposed methods is demonstrated through numerical experiments, data from sports analytics, and breast cancer data.

We present a tunable magnetoelastic lattice with a multistable onsite potential, focusing on a tristable potential. Through experimental and numerical analysis, we verify the existence of three types of transition waves with distinct amplitudes and velocities. Additionally, we establish the presence of a scaling law that elucidates various characteristics of these transition waves. By manipulating the onsite potential, we investigate the collision dynamics of two transition waves within the system. In chains featuring an asymmetric potential well, the collision of similar transition waves leads to the remote nucleation of a new phase. In chains with a symmetric potential well, the collision of dissimilar transition waves results in the formation of a stationary domain wall

Traffic congestion has significant impacts on both the economy and the environment. Measures of Effectiveness (MOEs) have long been the standard for evaluating the level of service and operational efficiency of traffic intersections. However, the scarcity of traditional high-resolution loop detector data (ATSPM) presents challenges in accurately measuring MOEs or capturing the intricate temporospatial characteristics inherent in urban intersection traffic. In response to this challenge, we have introduced the Multi-Task Deep Learning Digital Twin (MTDT) as a solution for multifaceted and precise intersection traffic flow simulation. MTDT enables accurate, fine-grained estimation of loop detector waveform time series for each lane of movement, alongside successful estimation of several MOEs for each lane group associated with a traffic phase concurrently and for all approaches of an arbitrary urban intersection. Unlike existing deep learning methodologies, MTDT distinguishes itself through its adaptability to local temporal and spatial features, such as signal timing plans, intersection topology, driving behaviors, and turning movement counts. While maintaining a straightforward design, our model emphasizes the advantages of multi-task learning in traffic modeling. By consolidating the learning process across multiple tasks, MTDT demonstrates reduced overfitting, increased efficiency, and enhanced effectiveness by sharing representations learned by different tasks. Furthermore, our approach facilitates sequential computation and lends itself to complete parallelization through GPU implementation. This not only streamlines the computational process but also enhances scalability and performance., Comment: 8 pages, 2 figures, 4 tables