Search

Efficient encoding and representation of large 3D molecular structures with high fidelity is critical for biomolecular design applications. Despite this, many representation learning approaches restrict themselves to modeling smaller systems or use coarse-grained approximations of the systems, for example modeling proteins at the resolution of amino acid residues rather than at the level of individual atoms. To address this, we develop quantized auto-encoders that learn atom-level tokenizations of complete proteins, RNA and small molecule structures with reconstruction accuracies well below 1 Angstrom. We demonstrate that a simple Mamba state space model architecture is efficient compared to an SE(3)-invariant IPA architecture, reaches competitive accuracies and can scale to systems with almost 100,000 atoms. The learned structure tokens of bio2token may serve as the input for all-atom generative models in the future.

The integration of distributed energy resources (DERs) into wholesale energy markets can greatly enhance grid flexibility, improve market efficiency, and contribute to a more sustainable energy future. As DERs -- such as solar PV panels and energy storage -- proliferate, effective mechanisms are needed to ensure that small prosumers can participate meaningfully in these markets. We study a wholesale market model featuring multiple DER aggregators, each controlling a portfolio of DER resources and bidding into the market on behalf of the DER asset owners. The key of our approach lies in recognizing the repeated nature of market interactions the ability of participants to learn and adapt over time. Specifically, Aggregators repeatedly interact with each other and with other suppliers in the wholesale market, collectively shaping wholesale electricity prices (aka the locational marginal prices (LMPs)). We model this multi-agent interaction using a mean-field game (MFG), which uses market information -- reflecting the average behavior of market participants -- to enable each aggregator to predict long-term LMP trends and make informed decisions. For each aggregator, because they control the DERs within their portfolio under certain contract structures, we employ a mean-field control (MFC) approach (as opposed to a MFG) to learn an optimal policy that maximizes the total rewards of the DERs under their management. We also propose a reinforcement learning (RL)-based method to help each agent learn optimal strategies within the MFG framework, enhancing their ability to adapt to market conditions and uncertainties. Numerical simulations show that LMPs quickly reach a steady state in the hybrid mean-field approach. Furthermore, our results demonstrate that the combination of energy storage and mean-field learning significantly reduces price volatility compared to scenarios without storage.

The environments where individuals live can present diverse navigation challenges, resulting in varying navigation abilities and strategies. Inspired by differing urban layouts and the Dual Solutions Paradigm test used for human navigators, we developed a simulated navigation environment to train deep reinforcement learning agents in a shortcut usage task. We modulated the frequency of exposure to a shortcut and navigation cue, leading to the development of artificial agents with differing abilities. We examined the encoded representations in artificial neural networks driving these agents, revealing intricate dynamics in representation learning, and correlated them with shortcut use preferences. Furthermore, we demonstrated methods to analyze representations across a population of nodes, which proved effective in finding patterns in what would otherwise be noisy single-node data. These techniques may also have broader applications in studying neural activity. From our observations in representation learning dynamics, we propose insights for human navigation learning, emphasizing the importance of navigation challenges in developing strong landmark knowledge over repeated exposures to landmarks alone.

We reassess the Knowledge Neuron (KN) Thesis: an interpretation of the mechanism underlying the ability of large language models to recall facts from a training corpus. This nascent thesis proposes that facts are recalled from the training corpus through the MLP weights in a manner resembling key-value memory, implying in effect that "knowledge" is stored in the network. Furthermore, by modifying the MLP modules, one can control the language model's generation of factual information. The plausibility of the KN thesis has been demonstrated by the success of KN-inspired model editing methods (Dai et al., 2022; Meng et al., 2022). We find that this thesis is, at best, an oversimplification. Not only have we found that we can edit the expression of certain linguistic phenomena using the same model editing methods but, through a more comprehensive evaluation, we have found that the KN thesis does not adequately explain the process of factual expression. While it is possible to argue that the MLP weights store complex patterns that are interpretable both syntactically and semantically, these patterns do not constitute "knowledge." To gain a more comprehensive understanding of the knowledge representation process, we must look beyond the MLP weights and explore recent models' complex layer structures and attention mechanisms., Comment: ICLR 2024 (Spotlight)

The White House Executive Order on Artificial Intelligence highlights the risks of large language models (LLMs) empowering malicious actors in developing biological, cyber, and chemical weapons. To measure these risks of malicious use, government institutions and major AI labs are developing evaluations for hazardous capabilities in LLMs. However, current evaluations are private, preventing further research into mitigating risk. Furthermore, they focus on only a few, highly specific pathways for malicious use. To fill these gaps, we publicly release the Weapons of Mass Destruction Proxy (WMDP) benchmark, a dataset of 3,668 multiple-choice questions that serve as a proxy measurement of hazardous knowledge in biosecurity, cybersecurity, and chemical security. WMDP was developed by a consortium of academics and technical consultants, and was stringently filtered to eliminate sensitive information prior to public release. WMDP serves two roles: first, as an evaluation for hazardous knowledge in LLMs, and second, as a benchmark for unlearning methods to remove such hazardous knowledge. To guide progress on unlearning, we develop RMU, a state-of-the-art unlearning method based on controlling model representations. RMU reduces model performance on WMDP while maintaining general capabilities in areas such as biology and computer science, suggesting that unlearning may be a concrete path towards reducing malicious use from LLMs. We release our benchmark and code publicly at https://wmdp.ai, Comment: See the project page at https://wmdp.ai

Utilizing distributed renewable and energy storage resources in local distribution networks via peer-to-peer (P2P) energy trading has long been touted as a solution to improve energy systems' resilience and sustainability. Consumers and prosumers (those who have energy generation resources), however, do not have the expertise to engage in repeated P2P trading, and the zero-marginal costs of renewables present challenges in determining fair market prices. To address these issues, we propose multi-agent reinforcement learning (MARL) frameworks to help automate consumers' bidding and management of their solar PV and energy storage resources, under a specific P2P clearing mechanism that utilizes the so-called supply-demand ratio. In addition, we show how the MARL frameworks can integrate physical network constraints to realize voltage control, hence ensuring physical feasibility of the P2P energy trading and paving way for real-world implementations.

This note aims at investigating two different situations where the classical general relativistic dynamics competes with the evolution driven by Hawking evaporation. We focus, in particular, on binary systems of black holes emitting gravitational waves and gravitons, and on the cosmological evolution when black holes are immersed in their own radiation bath. Several non-trivial features are underlined in both cases., Comment: 16 pages, 14 figures

Recent advancements in large language models (LLMs) have transformed the field of question answering (QA). However, evaluating LLMs in the medical field is challenging due to the lack of standardized and comprehensive datasets. To address this gap, we introduce CMExam, sourced from the Chinese National Medical Licensing Examination. CMExam consists of 60K+ multiple-choice questions for standardized and objective evaluations, as well as solution explanations for model reasoning evaluation in an open-ended manner. For in-depth analyses of LLMs, we invited medical professionals to label five additional question-wise annotations, including disease groups, clinical departments, medical disciplines, areas of competency, and question difficulty levels. Alongside the dataset, we further conducted thorough experiments with representative LLMs and QA algorithms on CMExam. The results show that GPT-4 had the best accuracy of 61.6% and a weighted F1 score of 0.617. These results highlight a great disparity when compared to human accuracy, which stood at 71.6%. For explanation tasks, while LLMs could generate relevant reasoning and demonstrate improved performance after finetuning, they fall short of a desired standard, indicating ample room for improvement. To the best of our knowledge, CMExam is the first Chinese medical exam dataset to provide comprehensive medical annotations. The experiments and findings of LLM evaluation also provide valuable insights into the challenges and potential solutions in developing Chinese medical QA systems and LLM evaluation pipelines. The dataset and relevant code are available at https://github.com/williamliujl/CMExam., Comment: Accepted by NeurIPS 2023 Datasets and Benchmarks Track

Navigation is a complex skill with a long history of research in animals and humans. In this work, we simulate the Morris Water Maze in 2D to train deep reinforcement learning agents. We perform automatic classification of navigation strategies, analyze the distribution of strategies used by artificial agents, and compare them with experimental data to show similar learning dynamics as those seen in humans and rodents. We develop environment-specific auxiliary tasks and examine factors affecting their usefulness. We suggest that the most beneficial tasks are potentially more biologically feasible for real agents to use. Lastly, we explore the development of internal representations in the activations of artificial agent neural networks. These representations resemble place cells and head-direction cells found in mouse brains, and their presence has correlation to the navigation strategies that artificial agents employ., Comment: 28 pages, 16 figures

Under the Autonomous Mobile Clinics (AMCs) initiative, we are developing, open sourcing, and standardizing health AI technologies to enable healthcare access in least developed countries (LDCs). We deem AMCs as the next generation of health care delivery platforms, whereas health AI engines are applications on these platforms, similar to how various applications expand the usage scenarios of smart phones. Facing the recent global monkeypox outbreak, in this article, we introduce AICOM-MP, an AI-based monkeypox detector specially aiming for handling images taken from resource-constrained devices. Compared to existing AI-based monkeypox detectors, AICOM-MP has achieved state-of-the-art (SOTA) performance. We have hosted AICOM-MP as a web service to allow universal access to monkeypox screening technology. We have also open sourced both the source code and the dataset of AICOM-MP to allow health AI professionals to integrate AICOM-MP into their services. Also, through the AICOM-MP project, we have generalized a methodology of developing health AI technologies for AMCs to allow universal access even in resource-constrained environments.

Distributed renewable resources owned by prosumers can be an effective way of fortifying grid resilience and enhancing sustainability. However, prosumers serve their own interests and their objectives are unlikely to align with that of society. This paper develops a bilevel model to study the optimal design of retail electricity tariffs considering the balance between economic efficiency and energy equity. The retail tariff entails a fixed charge and a volumetric charge tied to electricity usage to recover utilities' fixed costs. We analyze solution properties of the bilevel problem and prove an optimal rate design, which is to use fixed charges to recover fixed costs and to balance energy equity among different income groups. This suggests that programs similar to CARE (California Alternative Rate of Energy), which offer lower retail rates to low-income households, are unlikely to be efficient, even if they are politically appealing.

The chromatic polynomial of a graph $G$, denoted $P(G,m)$, is equal to the number of proper $m$-colorings of $G$ for each $m \in \mathbb{N}$. In 1990, Kostochka and Sidorenko introduced the list color function of graph $G$, denoted $P_{\ell}(G,m)$, which is a list analogue of the chromatic polynomial. The list color function threshold of $G$, denoted $\tau(G)$, is the smallest $k \geq \chi(G)$ such that $P_{\ell}(G,m) = P(G,m)$ whenever $m \geq k$. It is known that for every graph $G$, $\tau(G)$ is finite, and in fact, $\tau(G) \leq (|E(G)|-1)/\ln(1+ \sqrt{2}) + 1$. It is also known that when $G$ is a cycle or chordal graph, $G$ is enumeratively chromatic-choosable which means $\tau(G) = \chi(G)$. A recent paper of Kaul et al. suggests that understanding the list color function threshold of complete bipartite graphs is essential to the study of the extremal behavior of $\tau$. In this paper we show that for any $n \geq 2$, $\tau(K_{2,n}) \leq \lceil (n+2.05)/1.24 \rceil$ which gives an improvement on the general upper bound for $\tau(G)$ when $G = K_{2,n}$. We also develop additional tools that allow us to show that $\tau(K_{2,3}) = \chi(K_{2,3})$ and $\tau(K_{2,4}) = \tau(K_{2,5}) = 3$., Comment: 30 pages

The goal of this work is to perform 3D reconstruction and novel view synthesis from data captured by scanning platforms commonly deployed for world mapping in urban outdoor environments (e.g., Street View). Given a sequence of posed RGB images and lidar sweeps acquired by cameras and scanners moving through an outdoor scene, we produce a model from which 3D surfaces can be extracted and novel RGB images can be synthesized. Our approach extends Neural Radiance Fields, which has been demonstrated to synthesize realistic novel images for small scenes in controlled settings, with new methods for leveraging asynchronously captured lidar data, for addressing exposure variation between captured images, and for leveraging predicted image segmentations to supervise densities on rays pointing at the sky. Each of these three extensions provides significant performance improvements in experiments on Street View data. Our system produces state-of-the-art 3D surface reconstructions and synthesizes higher quality novel views in comparison to both traditional methods (e.g.~COLMAP) and recent neural representations (e.g.~Mip-NeRF)., Comment: Project: https://urban-radiance-fields.github.io/

Modern electric power systems have witnessed rapidly increasing penetration of renewable energy, storage, electrical vehicles and various demand response resources. The electric infrastructure planning is thus facing more challenges due to the variability and uncertainties arising from the diverse new resources. This study aims to develop a multistage and multiscale stochastic mixed integer programming (MM-SMIP) model to capture both the coarse-temporal-scale uncertainties, such as investment cost and long-run demand stochasticity, and fine-temporal-scale uncertainties, such as hourly renewable energy output and electricity demand uncertainties, for the power system capacity expansion problem. To be applied to a real power system, the resulting model will lead to extremely large-scale mixed integer programming problems, which suffer not only the well-known curse of dimensionality, but also computational difficulties with a vast number of integer variables at each stage. In addressing such challenges associated with the MM-SMIP model, we propose a nested cross decomposition algorithm that consists of two layers of decomposition, that is, the Dantzig-Wolfe decomposition and L-shaped decomposition. The algorithm exhibits promising computational performance under our numerical study, and is especially amenable to parallel computing, which will also be demonstrated through the computational results.

A commonly observed failure mode of Neural Radiance Field (NeRF) is fitting incorrect geometries when given an insufficient number of input views. One potential reason is that standard volumetric rendering does not enforce the constraint that most of a scene's geometry consist of empty space and opaque surfaces. We formalize the above assumption through DS-NeRF (Depth-supervised Neural Radiance Fields), a loss for learning radiance fields that takes advantage of readily-available depth supervision. We leverage the fact that current NeRF pipelines require images with known camera poses that are typically estimated by running structure-from-motion (SFM). Crucially, SFM also produces sparse 3D points that can be used as "free" depth supervision during training: we add a loss to encourage the distribution of a ray's terminating depth matches a given 3D keypoint, incorporating depth uncertainty. DS-NeRF can render better images given fewer training views while training 2-3x faster. Further, we show that our loss is compatible with other recently proposed NeRF methods, demonstrating that depth is a cheap and easily digestible supervisory signal. And finally, we find that DS-NeRF can support other types of depth supervision such as scanned depth sensors and RGB-D reconstruction outputs., Comment: Project page: http://www.cs.cmu.edu/~dsnerf/ GitHub: https://github.com/dunbar12138/DSNeRF

Tensor kernels in machine learning (ML) often correspond to pure mathematical expressions, making term rewriting an attractive strategy for optimization and mapping to specialized hardware accelerators. However, existing ML intermediate representations (IRs) tend to either be \textit{pure but high-level}, making low-level rewrites to hardware targets inexpressible, or \textit{low-level but impure}, hampering the use of term rewriting altogether. This paper introduces Glenside, a pure IR whose core abstraction -- the \textit{access pattern} -- enables low-level, layout-aware, hardware-centric program rewrites. We demonstrate how term rewriting in Glenside can be used to map program fragments to hardware accelerator invocations and automatically discover classic data layout transformations like \texttt{im2col}. Glenside establishes a new foundation for exploring further term rewriting techniques in optimizing low-level tensor programs., Comment: To be published at MAPS 2021

We present a framework for automatically reconfiguring images of street scenes by populating, depopulating, or repopulating them with objects such as pedestrians or vehicles. Applications of this method include anonymizing images to enhance privacy, generating data augmentations for perception tasks like autonomous driving, and composing scenes to achieve a certain ambiance, such as empty streets in the early morning. At a technical level, our work has three primary contributions: (1) a method for clearing images of objects, (2) a method for estimating sun direction from a single image, and (3) a way to compose objects in scenes that respects scene geometry and illumination. Each component is learned from data with minimal ground truth annotations, by making creative use of large-numbers of short image bursts of street scenes. We demonstrate convincing results on a range of street scenes and illustrate potential applications., Comment: CVPR 2021

Quantization is a key technique to reduce the resource requirement and improve the performance of neural network deployment. However, different hardware backends such as x86 CPU, NVIDIA GPU, ARM CPU, and accelerators may demand different implementations for quantized networks. This diversity calls for specialized post-training quantization pipelines to built for each hardware target, an engineering effort that is often too large for developers to keep up with. We tackle this problem with an automated post-training quantization framework called HAGO. HAGO provides a set of general quantization graph transformations based on a user-defined hardware specification and implements a search mechanism to find the optimal quantization strategy while satisfying hardware constraints for any model. We observe that HAGO achieves speedups of 2.09x, 1.97x, and 2.48x on Intel Xeon Cascade Lake CPUs, NVIDIA Tesla T4 GPUs, ARM Cortex-A CPUs on Raspberry Pi4 relative to full precision respectively, while maintaining the highest reported post-training quantization accuracy in each case.

This paper first proposes an N-block PCPM algorithm to solve N-block convex optimization problems with both linear and nonlinear constraints, with global convergence established. A linear convergence rate under the strong second-order conditions for optimality is observed in the numerical experiments. Next, for a starting point, an asynchronous N-block PCPM algorithm is proposed to solve linearly constrained N-block convex optimization problems. The numerical results demonstrate the sub-linear convergence rate under the bounded delay assumption, as well as the faster convergence with more short-time iterations than a synchronous iterative scheme.

This paper applies the N-block PCPM algorithm to solve multi-scale multi-stage stochastic programs, with the application to electricity capacity expansion models. Numerical results show that the proposed simplified N-block PCPM algorithm, along with the hybrid decomposition method, exhibits much better scalability for solving the resulting deterministic, large-scale block-separable optimization problem when compared with the ADMM algorithm and the PHA algorithm. The superiority of the algorithm's scalability is attributed to the two key features of the algorithm design: first, the proposed hybrid scenario-node-realization decomposition method with extended nonanticipativity constraints can decompose the original problem under various uncertainties of different temporal scales; second, when applying the N-block PCPM algorithm to solve the resulting deterministic, large-scale N-block convex optimization problem, the technique of orthogonal projection we exploit greatly simplifies the iteration steps and reduce the communication overhead among all computing units, which also contributes to the efficiency of the algorithm., Comment: The paper needs some improvements

Reduction in the cost of Network Cameras along with a rise in connectivity enables entities all around the world to deploy vast arrays of camera networks. Network cameras offer real-time visual data that can be used for studying traffic patterns, emergency response, security, and other applications. Although many sources of Network Camera data are available, collecting the data remains difficult due to variations in programming interface and website structures. Previous solutions rely on manually parsing the target website, taking many hours to complete. We create a general and automated solution for aggregating Network Camera data spread across thousands of uniquely structured web pages. We analyze heterogeneous web page structures and identify common characteristics among 73 sample Network Camera websites (each website has multiple web pages). These characteristics are then used to build an automated camera discovery module that crawls and aggregates Network Camera data. Our system successfully extracts 57,364 Network Cameras from 237,257 unique web pages., Comment: This paper has been accepted by ACM Transactions on Internet Technology

The implementation of electricity markets based on locational marginal pricing in a multi-settlement process has allowed wholesale competition, with pricing mechanisms that incentivize the optimal allocation of generation, transmission, and demand-response resources. While efficiency and reliability gains have been achieved in the US at a regional level, the lack of adequate inter-regional coordination mechanisms has limited broader gains. The shortcomings of cross-border coordination of electricity markets become more apparent as the industry transitions towards higher penetration of renewable resources, which tend to be concentrated in certain regions of the country, usually distant from load pockets. In addition to allowing market participants to respond to economic signals beyond the market region where their assets reside, a coordinated market solution would allow the extension of market mechanisms such as financial transmission rights and capacity contracts to cover transactions across borders; thereby, reducing price uncertainty around the investment in new generation resources. This paper presents a distributed solution to the unit commitment problem that allows for full coordination of the market clearing process across interconnected electricity markets.

DP-coloring (also called correspondence coloring) is a generalization of list coloring that has been widely studied in recent years after its introduction by Dvo\v{r}\'{a}k and Postle in 2015. As the analogue of the chromatic polynomial $P(G,m)$, the DP color function of a graph $G$, denoted $P_{DP}(G,m)$, counts the minimum number of DP-colorings over all possible $m$-fold covers. It is known that, unlike the list color function $P_{\ell}(G,m)$, for any $g \geq 3$ there exists a graph $G$ with girth $g$ such that $P_{DP}(G,m) < P(G,m)$ when $m$ is sufficiently large. Thus, two fundamental open questions regarding the DP color function are: (i) for which $G$ does there exist an $N \in \mathbb{N}$ such that $P_{DP}(G,m) = P(G,m)$ whenever $m \geq N$, (ii) Given a graph $G$ does there always exist an $N \in \mathbb{N}$ and a polynomial $p(m)$ such that $P_{DP}(G,m) = p(m)$ whenever $m \geq N$? In this paper we give exact formulas for the DP color function of a Theta graph based on the parity of its path lengths. This gives an explicit answer, including the formulas for the polynomials that are not the chromatic polynomial, to both the questions above for Theta graphs. We extend this result to Generalized Theta graphs by characterizing the exact parity condition that ensures the DP color function eventually equals the chromatic polynomial. To answer the second question for Generalized Theta graphs, we confirm it for the larger class of graphs with a feedback vertex set of size one., Comment: 21 pages. arXiv admin note: text overlap with arXiv:2009.08242, arXiv:1904.07697

We introduce the problem of perpetual view generation - long-range generation of novel views corresponding to an arbitrarily long camera trajectory given a single image. This is a challenging problem that goes far beyond the capabilities of current view synthesis methods, which quickly degenerate when presented with large camera motions. Methods for video generation also have limited ability to produce long sequences and are often agnostic to scene geometry. We take a hybrid approach that integrates both geometry and image synthesis in an iterative `\emph{render}, \emph{refine} and \emph{repeat}' framework, allowing for long-range generation that cover large distances after hundreds of frames. Our approach can be trained from a set of monocular video sequences. We propose a dataset of aerial footage of coastal scenes, and compare our method with recent view synthesis and conditional video generation baselines, showing that it can generate plausible scenes for much longer time horizons over large camera trajectories compared to existing methods. Project page at https://infinite-nature.github.io/., Comment: ICCV 2021 (oral); Project page: https://infinite-nature.github.io/; Video: https://www.youtube.com/watch?v=oXUf6anNAtc

We propose a learning-based framework for disentangling outdoor scenes into temporally-varying illumination and permanent scene factors. Inspired by the classic intrinsic image decomposition, our learning signal builds upon two insights: 1) combining the disentangled factors should reconstruct the original image, and 2) the permanent factors should stay constant across multiple temporal samples of the same scene. To facilitate training, we assemble a city-scale dataset of outdoor timelapse imagery from Google Street View, where the same locations are captured repeatedly through time. This data represents an unprecedented scale of spatio-temporal outdoor imagery. We show that our learned disentangled factors can be used to manipulate novel images in realistic ways, such as changing lighting effects and scene geometry. Please visit factorize-a-city.github.io for animated results., Comment: ECCV 2020 (Spotlight). Supplemental Material attached

Distributed energy resources (DERs), such as rooftop solar panels, are growing rapidly and are reshaping power systems. To promote DERs, feed-in-tariff (FIT) is usually adopted by utilities to pay DER owners certain fixed rates for supplying energy to the grid. An alternative to FIT is a market based approach; i.e., consumers and DER owners trade energy in an auction-based peer-to-peer (P2P) market, and the rates are determined by a market clearing process. However, the complexities in sucha market and agents' bounded rationality may invalidate many well-established theories on auction design and hinder market development. To address this issue, we propose an automated bidding framework in a repeated auction based on multi-armed bandit learning, which aims to minimize each bidder's cumulative regret. Numerical results indicate convergence of such a multi-agent learning game to a steady-state. For comparison purpose, we apply the framework to three different auction designs to realize a P2P market.