Search

Implicit neural representations (INRs) have demonstrated success in a variety of applications, including inverse problems and neural rendering. An INR is typically trained to capture one signal of interest, resulting in learned neural features that are highly attuned to that signal. Assumed to be less generalizable, we explore the aspect of transferability of such learned neural features for fitting similar signals. We introduce a new INR training framework, STRAINER that learns transferrable features for fitting INRs to new signals from a given distribution, faster and with better reconstruction quality. Owing to the sequential layer-wise affine operations in an INR, we propose to learn transferable representations by sharing initial encoder layers across multiple INRs with independent decoder layers. At test time, the learned encoder representations are transferred as initialization for an otherwise randomly initialized INR. We find STRAINER to yield extremely powerful initialization for fitting images from the same domain and allow for $\approx +10dB$ gain in signal quality early on compared to an untrained INR itself. STRAINER also provides a simple way to encode data-driven priors in INRs. We evaluate STRAINER on multiple in-domain and out-of-domain signal fitting tasks and inverse problems and further provide detailed analysis and discussion on the transferability of STRAINER's features. Our demo can be accessed at https://kushalvyas.github.io/strainer.html ., Comment: Project Website: https://kushalvyas.github.io/strainer.html

Graph signals are functions of the underlying graph. When the edge-weight between a pair of nodes is high, the corresponding signals generally have a higher correlation. As a result, the signals can be represented in terms of a graph-based generative model. The question then arises whether measurements can be obtained on a few nodes and whether the correlation structure between the signals can be used to reconstruct the graph signal on the remaining nodes. We show that node subsampling is always possible for graph signals obtained through a generative model. Further, a method to determine the number of nodes to select is proposed based on the tolerable error. A correlation-based fast greedy algorithm is developed for selecting the nodes. Finally, we verify the proposed method on different deterministic and random graphs, and show that near-perfect reconstruction is possible with node subsampling.

A great advantage of legged robots is their ability to operate on particularly difficult and obstructed terrain, which demands dynamic, robust, and precise movements. The study of obstacle courses provides invaluable insights into the challenges legged robots face, offering a controlled environment to assess and enhance their capabilities. Traversing it with a one-legged hopper introduces intricate challenges, such as planning over contacts and dealing with flight phases, which necessitates a sophisticated controller. A novel model predictive parkour controller is introduced, that finds an optimal path through a real-time changing obstacle course with mixed integer motion planning. The execution of this optimized path is then achieved through a state machine employing a PD control scheme with feedforward torques, ensuring robust and accurate performance., Comment: Published in: IEEE Robotics and Automation Letters

DarkSide-20k (DS-20k) is a dark matter detection experiment under construction at the Laboratori Nazionali del Gran Sasso (LNGS) in Italy. It utilises ~100 t of low radioactivity argon from an underground source (UAr) in its inner detector, with half serving as target in a dual-phase time projection chamber (TPC). The UAr cryogenics system must maintain stable thermodynamic conditions throughout the experiment's lifetime of >10 years. Continuous removal of impurities and radon from the UAr is essential for maximising signal yield and mitigating background. We are developing an efficient and powerful cryogenics system with a gas purification loop with a target circulation rate of 1000 slpm. Central to its design is a condenser operated with liquid nitrogen which is paired with a gas heat exchanger cascade, delivering a combined cooling power of >8 kW. Here we present the design choices in view of the DS-20k requirements, in particular the condenser's working principle and the cooling control, and we show test results obtained with a dedicated benchmarking platform at CERN and LNGS. We find that the thermal efficiency of the recirculation loop, defined in terms of nitrogen consumption per argon flow rate, is 95 % and the pressure in the test cryostat can be maintained within $\pm$(0.1-0.2) mbar. We further detail a 5-day cool-down procedure of the test cryostat, maintaining a cooling rate typically within -2 K/h, as required for the DS-20k inner detector. Additionally, we assess the circuit's flow resistance, and the heat transfer capabilities of two heat exchanger geometries for argon phase change, used to provide gas for recirculation. We conclude by discussing how our findings influence the finalisation of the system design, including necessary modifications to meet requirements and ongoing testing activities., Comment: 45 pages, 24 figures

Continuous glucose monitoring (CGM) devices provide real-time glucose monitoring and timely alerts for glycemic excursions, improving glycemic control among patients with diabetes. However, identifying rare events like hypoglycemia and hyperglycemia remain challenging due to their infrequency. Moreover, limited access to sensitive patient data hampers the development of robust machine learning models. Our objective is to accurately predict glycemic excursions while addressing data privacy concerns. To tackle excursion prediction, we propose a novel Hypo-Hyper (HH) loss function, which significantly improves performance in the glycemic excursion regions. The HH loss function demonstrates a 46% improvement over mean-squared error (MSE) loss across 125 patients. To address privacy concerns, we propose FedGlu, a machine learning model trained in a federated learning (FL) framework. FL allows collaborative learning without sharing sensitive data by training models locally and sharing only model parameters across other patients. FedGlu achieves a 35% superior glycemic excursion detection rate compared to local models. This improvement translates to enhanced performance in predicting both, hypoglycemia and hyperglycemia, for 105 out of 125 patients. These results underscore the effectiveness of the proposed HH loss function in augmenting the predictive capabilities of glucose predictions. Moreover, implementing models within a federated learning framework not only ensures better predictive capabilities but also safeguards sensitive data concurrently.

This work presents the design and development of IIT Bombay Racing's Formula Student style autonomous racecar algorithm capable of running at the racing events of Formula Student-AI, held in the UK. The car employs a cutting-edge sensor suite of the compute unit NVIDIA Jetson Orin AGX, 2 ZED2i stereo cameras, 1 Velodyne Puck VLP16 LiDAR and SBG Systems Ellipse N GNSS/INS IMU. It features deep learning algorithms and control systems to navigate complex tracks and execute maneuvers without any human intervention. The design process involved extensive simulations and testing to optimize the vehicle's performance and ensure its safety. The algorithms have been tested on a small scale, in-house manufactured 4-wheeled robot and on simulation software. The results obtained for testing various algorithms in perception, simultaneous localization and mapping, path planning and controls have been detailed., Comment: 8 pages, 19 figures

This paper investigates the threat of backdoors in Deep Reinforcement Learning (DRL) agent policies and proposes a novel method for their detection at runtime. Our study focuses on elusive in-distribution backdoor triggers. Such triggers are designed to induce a deviation in the behaviour of a backdoored agent while blending into the expected data distribution to evade detection. Through experiments conducted in the Atari Breakout environment, we demonstrate the limitations of current sanitisation methods when faced with such triggers and investigate why they present a challenging defence problem. We then evaluate the hypothesis that backdoor triggers might be easier to detect in the neural activation space of the DRL agent's policy network. Our statistical analysis shows that indeed the activation patterns in the agent's policy network are distinct in the presence of a trigger, regardless of how well the trigger is concealed in the environment. Based on this, we propose a new defence approach that uses a classifier trained on clean environment samples and detects abnormal activations. Our results show that even lightweight classifiers can effectively prevent malicious actions with considerable accuracy, indicating the potential of this research direction even against sophisticated adversaries., Comment: 11 Pages, 12 figures

We construct 2-query, quasi-linear size probabilistically checkable proofs (PCPs) with arbitrarily small constant soundness, improving upon Dinur's 2-query quasi-linear size PCPs with soundness $1-\Omega(1)$. As an immediate corollary, we get that under the exponential time hypothesis, for all $\epsilon>0$ no approximation algorithm for $3$-SAT can obtain an approximation ratio of $7/8+\epsilon$ in time $2^{n/\log^C n}$, where $C$ is a constant depending on $\epsilon$. Our result builds on a recent line of independent works by Bafna, Lifshitz and Minzer, and Dikstein, Dinur and Lubotzky, that showed the existence of linear size direct product testers with small soundness. The main new ingredient in our proof is a technique that embeds a given 2-CSP into a 2-CSP on a prescribed graph, provided that the latter is a graph underlying a sufficiently good high-dimensional expander (HDX). We achieve this by establishing a novel connection between PCPs and fault-tolerant distributed computing, more precisely, to the almost-everywhere reliable transmission problem introduced by Dwork, Peleg, Pippenger and Upfal (1986). We instantiate this connection by showing that graphs underlying HDXs admit routing protocols that are tolerant to adversarial edge corruptions, also improving upon the state of the art constructions of sparse edge-fault-tolerant networks in the process. Our PCP construction requires variants of the aforementioned direct product testers with poly-logarithmic degree. The existence and constructability of these variants is shown in an appendix by Zhiwei Yun., Comment: Appendix D by Zhiwei Yun showing that variants of the Chapman-Lubotzky complexes can be constructed with $q=\text{polylog}(n)$

Training language models becomes increasingly expensive with scale, prompting numerous attempts to improve optimization efficiency. Despite these efforts, the Adam optimizer remains the most widely used, due to a prevailing view that it is the most effective approach. We aim to compare several optimization algorithms, including SGD, Adafactor, Adam, and Lion, in the context of autoregressive language modeling across a range of model sizes, hyperparameters, and architecture variants. Our findings indicate that, except for SGD, these algorithms all perform comparably both in their optimal performance and also in terms of how they fare across a wide range of hyperparameter choices. Our results suggest to practitioners that the choice of optimizer can be guided by practical considerations like memory constraints and ease of implementation, as no single algorithm emerged as a clear winner in terms of performance or stability to hyperparameter misspecification. Given our findings, we further dissect these approaches, examining two simplified versions of Adam: a) signed momentum (Signum) which we see recovers both the performance and hyperparameter stability of Adam and b) Adalayer, a layerwise variant of Adam which we introduce to study Adam's preconditioning. Examining Adalayer leads us to the conclusion that the largest impact of Adam's preconditioning is restricted to the last layer and LayerNorm parameters, and, perhaps surprisingly, the remaining layers can be trained with SGD.

The dual-phase liquid argon time projection chamber is presently one of the leading technologies to search for dark matter particles with masses below 10 GeV/c$^2$. This was demonstrated by the DarkSide-50 experiment with approximately 50 kg of low-radioactivity liquid argon as target material. The next generation experiment DarkSide-20k, currently under construction, will use 1,000 times more argon and is expected to start operation in 2027. Based on the DarkSide-50 experience, here we assess the DarkSide-20k sensitivity to models predicting light dark matter particles, including Weakly Interacting Massive Particles (WIMPs) and sub-GeV/c$^2$ particles interacting with electrons in argon atoms. With one year of data, a sensitivity improvement to dark matter interaction cross-sections by at least one order of magnitude with respect to DarkSide-50 is expected for all these models. A sensitivity to WIMP--nucleon interaction cross-sections below $1\times10^{-42}$ cm$^2$ is achievable for WIMP masses above 800 MeV/c$^2$. With 10 years exposure, the neutrino fog can be reached for WIMP masses around 5 GeV/c$^2$., Comment: 13 pages, 4 figures, supplementary material (4 figures)

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.

We focus on Text-to-SQL semantic parsing from the perspective of retrieval-augmented generation. Motivated by challenges related to the size of commercial database schemata and the deployability of business intelligence solutions, we propose $\text{ASTReS}$ that dynamically retrieves input database information and uses abstract syntax trees to select few-shot examples for in-context learning. Furthermore, we investigate the extent to which an in-parallel semantic parser can be leveraged for generating approximated versions of the expected SQL queries, to support our retrieval. We take this approach to the extreme--we adapt a model consisting of less than $500$M parameters, to act as an extremely efficient approximator, enhancing it with the ability to process schemata in a parallelised manner. We apply $\text{ASTReS}$ to monolingual and cross-lingual benchmarks for semantic parsing, showing improvements over state-of-the-art baselines. Comprehensive experiments highlight the contribution of modules involved in this retrieval-augmented generation setting, revealing interesting directions for future work., Comment: EMNLP 2024 Main

Shampoo, a second-order optimization algorithm which uses a Kronecker product preconditioner, has recently garnered increasing attention from the machine learning community. The preconditioner used by Shampoo can be viewed either as an approximation of the Gauss--Newton component of the Hessian or the covariance matrix of the gradients maintained by Adagrad. We provide an explicit and novel connection between the $\textit{optimal}$ Kronecker product approximation of these matrices and the approximation made by Shampoo. Our connection highlights a subtle but common misconception about Shampoo's approximation. In particular, the $\textit{square}$ of the approximation used by the Shampoo optimizer is equivalent to a single step of the power iteration algorithm for computing the aforementioned optimal Kronecker product approximation. Across a variety of datasets and architectures we empirically demonstrate that this is close to the optimal Kronecker product approximation. Additionally, for the Hessian approximation viewpoint, we empirically study the impact of various practical tricks to make Shampoo more computationally efficient (such as using the batch gradient and the empirical Fisher) on the quality of Hessian approximation.

To explore oceans on ice-covered moons in the solar system, energy-efficient Autonomous Underwater Vehicles (AUVs) with long ranges must cover enough distance to record and collect enough data. These usually underactuated vehicles are hard to control when performing tasks such as vertical docking or the inspection of vertical walls. This paper introduces a control strategy for DeepLeng to navigate in the ice-covered ocean of Jupiter's moon Europa and presents simulation results preceding a discussion on what is further needed for robust control during the mission., Comment: 7 pages, 8 figures

The knowledge of scintillation quenching of $\alpha$-particles plays a paramount role in understanding $\alpha$-induced backgrounds and improving the sensitivity of liquid argon-based direct detection of dark matter experiments. We performed a relative measurement of scintillation quenching in the MeV energy region using radioactive isotopes ($^{222}$Rn, $^{218}$Po and $^{214}$Po isotopes) present in trace amounts in the DEAP-3600 detector and quantified the uncertainty of extrapolating the quenching factor to the low-energy region., Comment: 9 pages, 6 figures (added 1 figure, revised 3 figures), 2 tables, revised sections 3, 4, 5. Accepted in Eur. Phys. J. C

As the scale of generative models continues to grow, efficient reuse and adaptation of pre-trained models have become crucial considerations. In this work, we propose Voicebox Adapter, a novel approach that integrates fine-grained conditions into a pre-trained Voicebox speech generation model using a cross-attention module. To ensure a smooth integration of newly added modules with pre-trained ones, we explore various efficient fine-tuning approaches. Our experiment shows that the LoRA with bias-tuning configuration yields the best performance, enhancing controllability without compromising speech quality. Across three fine-grained conditional generation tasks, we demonstrate the effectiveness and resource efficiency of Voicebox Adapter. Follow-up experiments further highlight the robustness of Voicebox Adapter across diverse data setups., Comment: Accepted by InterSpeech 2024

The classical game theory models rational players and proposes Nash equilibrium (NE) as the solution. However, real-world scenarios rarely feature rational players; instead, players make inconsistent and irrational decisions. Often, irrational players exhibit herding behaviour by simply following the majority. In this paper, we consider the mean-field game with $\alpha$-fraction of rational players and the rest being herding-irrational players. For such a game, we introduce a novel concept of equilibrium named $\alpha$-Rational NE (in short, $\alpha$-RNE). The $\alpha$-RNEs and their implications are extensively analyzed in the game with two actions. Due to herding-irrational players, new equilibria may arise, and some classical NEs may be deleted. The rational players are not harmed but benefit from the presence of irrational players. Notably, we demonstrate through examples that rational players leverage upon the herding behaviour of irrational players and may attain higher utility (under $\alpha$-RNE) than social optimal utility (in the classical setting). Interestingly, the irrational players may also benefit by not being rational. We observe that irrational players do not lose compared to some classical NEs for participation and bandwidth sharing games. More importantly, in bandwidth sharing game, irrational players receive utility that approaches the social optimal utility. Such examples indicate that it may sometimes be `rational' to be irrational., Comment: 8 pages, 3 figure, 1 table