Search

Imitation learning has demonstrated significant potential in performing high-precision manipulation tasks using visual feedback from cameras. However, it is common practice in imitation learning for cameras to be fixed in place, resulting in issues like occlusion and limited field of view. Furthermore, cameras are often placed in broad, general locations, without an effective viewpoint specific to the robot's task. In this work, we investigate the utility of active vision (AV) for imitation learning and manipulation, in which, in addition to the manipulation policy, the robot learns an AV policy from human demonstrations to dynamically change the robot's camera viewpoint to obtain better information about its environment and the given task. We introduce AV-ALOHA, a new bimanual teleoperation robot system with AV, an extension of the ALOHA 2 robot system, incorporating an additional 7-DoF robot arm that only carries a stereo camera and is solely tasked with finding the best viewpoint. This camera streams stereo video to an operator wearing a virtual reality (VR) headset, allowing the operator to control the camera pose using head and body movements. The system provides an immersive teleoperation experience, with bimanual first-person control, enabling the operator to dynamically explore and search the scene and simultaneously interact with the environment. We conduct imitation learning experiments of our system both in real-world and in simulation, across a variety of tasks that emphasize viewpoint planning. Our results demonstrate the effectiveness of human-guided AV for imitation learning, showing significant improvements over fixed cameras in tasks with limited visibility. Project website: https://soltanilara.github.io/av-aloha/, Comment: 6 pages, 4 figures

We present InterACT: Inter-dependency aware Action Chunking with Hierarchical Attention Transformers, a novel imitation learning framework for bimanual manipulation that integrates hierarchical attention to capture inter-dependencies between dual-arm joint states and visual inputs. InterACT consists of a Hierarchical Attention Encoder and a Multi-arm Decoder, both designed to enhance information aggregation and coordination. The encoder processes multi-modal inputs through segment-wise and cross-segment attention mechanisms, while the decoder leverages synchronization blocks to refine individual action predictions, providing the counterpart's prediction as context. Our experiments on a variety of simulated and real-world bimanual manipulation tasks demonstrate that InterACT significantly outperforms existing methods. Detailed ablation studies validate the contributions of key components of our work, including the impact of CLS tokens, cross-segment encoders, and synchronization blocks., Comment: Accepted at Conference on Robot Learning (CoRL) 2024

Modern generative models demonstrate impressive capabilities, likely stemming from an ability to identify and manipulate abstract concepts underlying their training data. However, fundamental questions remain: what determines the concepts a model learns, the order in which it learns them, and its ability to manipulate those concepts? To address these questions, we propose analyzing a model's learning dynamics via a framework we call the concept space, where each axis represents an independent concept underlying the data generating process. By characterizing learning dynamics in this space, we identify how the speed at which a concept is learned, and hence the order of concept learning, is controlled by properties of the data we term concept signal. Further, we observe moments of sudden turns in the direction of a model's learning dynamics in concept space. Surprisingly, these points precisely correspond to the emergence of hidden capabilities, i.e., where latent interventions show the model possesses the capability to manipulate a concept, but these capabilities cannot yet be elicited via naive input prompting. While our results focus on synthetically defined toy datasets, we hypothesize a general claim on emergence of hidden capabilities may hold: generative models possess latent capabilities that emerge suddenly and consistently during training, though a model might not exhibit these capabilities under naive input prompting., Comment: Preprint

This work characterizes equivariant polynomial functions from tuples of tensor inputs to tensor outputs. Loosely motivated by physics, we focus on equivariant functions with respect to the diagonal action of the orthogonal group on tensors. We show how to extend this characterization to other linear algebraic groups, including the Lorentz and symplectic groups. Our goal behind these characterizations is to define equivariant machine learning models. In particular, we focus on the sparse vector estimation problem. This problem has been broadly studied in the theoretical computer science literature, and explicit spectral methods, derived by techniques from sum-of-squares, can be shown to recover sparse vectors under certain assumptions. Our numerical results show that the proposed equivariant machine learning models can learn spectral methods that outperform the best theoretically known spectral methods in some regimes. The experiments also suggest that learned spectral methods can solve the problem in settings that have not yet been theoretically analyzed. This is an example of a promising direction in which theory can inform machine learning models and machine learning models could inform theory.

Creating multilingual task-oriented dialogue (TOD) agents is challenging due to the high cost of training data acquisition. Following the research trend of improving training data efficiency, we show for the first time, that in-context learning is sufficient to tackle multilingual TOD. To handle the challenging dialogue state tracking (DST) subtask, we break it down to simpler steps that are more compatible with in-context learning where only a handful of few-shot examples are used. We test our approach on the multilingual TOD dataset X-RiSAWOZ, which has 12 domains in Chinese, English, French, Korean, Hindi, and code-mixed Hindi-English. Our turn-by-turn DST accuracy on the 6 languages range from 55.6% to 80.3%, seemingly worse than the SOTA results from fine-tuned models that achieve from 60.7% to 82.8%; our BLEU scores in the response generation (RG) subtask are also significantly lower than SOTA. However, after manual evaluation of the validation set, we find that by correcting gold label errors and improving dataset annotation schema, GPT-4 with our prompts can achieve (1) 89.6%-96.8% accuracy in DST, and (2) more than 99% correct response generation across different languages. This leads us to conclude that current automatic metrics heavily underestimate the effectiveness of in-context learning.

Severe acute malnutrition (SAM, or severe wasting) and moderate acute malnutrition (MAM, or moderate wasting) affect 52 million children under five years of age around the globe. This systematic review, commissioned by the Humanitarian Evidence Programme and carried out by a research team from the University of Sheffield, represents the first attempt to apply systematic review methodology to establish the relationships between recovery and relapse and between default rates and repeated episodes of default or relapse in the management of acute malnutrition in children in humanitarian emergencies in low- and middle-income countries., The systematic review, together with corresponding executive summary and evidence brief, forms part of a series of humanitarian evidence syntheses and systematic reviews commissioned by the Humanitarian Evidence Programme. Other reports in the series review the evidence on interventions or approaches to mental health, child protection, market support and household food security, pastoralist livelihoods, shelter self-recovery and targeting in urban settings., The Humanitarian Evidence Programme is a partnership between Oxfam GB and the Feinstein International Center at the Friedman School of Nutrition Science and Policy, Tufts University. It is funded by the United Kingdom (UK) government’s Department for International Development (DFID) through the Humanitarian Innovation and Evidence Programme.

This protocol outlines plans for conducting a mixed-methods systematic review on acute malnutrition in humanitarian crises. The review will investigate the relationship between recovery/cure and relapse, and between relapse and default and/or return defaults/episodes of default in the management of acute malnutrition in children under five in humanitarian emergencies. The review will also explore the contexts in which acute malnutrition management programmes were implemented, in order to identify and describe how context influences relapse and default and/or return default/episodes. This review is funded through the Humanitarian Evidence Programme, a UK Aid-funded partnership between Oxfam and Feinstein International Center (FIC) at the Friedman School of Nutrition at Tufts University. The Humanitarian Evidence Programme aims to synthesize evidence in the humanitarian sector and communicate the findings to stakeholders, with the ultimate goal of improving humanitarian policy and practice.

Synthetic data generation has the potential to impact applications and domains with scarce data. However, before such data is used for sensitive tasks such as mental health, we need an understanding of how different demographics are represented in it. In our paper, we analyze the potential of producing synthetic data using GPT-3 by exploring the various stressors it attributes to different race and gender combinations, to provide insight for future researchers looking into using LLMs for data generation. Using GPT-3, we develop HEADROOM, a synthetic dataset of 3,120 posts about depression-triggering stressors, by controlling for race, gender, and time frame (before and after COVID-19). Using this dataset, we conduct semantic and lexical analyses to (1) identify the predominant stressors for each demographic group; and (2) compare our synthetic data to a human-generated dataset. We present the procedures to generate queries to develop depression data using GPT-3, and conduct analyzes to uncover the types of stressors it assigns to demographic groups, which could be used to test the limitations of LLMs for synthetic data generation for depression data. Our findings show that synthetic data mimics some of the human-generated data distribution for the predominant depression stressors across diverse demographics., Comment: 14 pages, 16 figures

We consider the embedding function $c_b(a)$ describing the problem of symplectically embedding an ellipsoid $E(1,a)$ into the smallest possible scaling by $\lambda>1$ of the polydisc $P(1,b)$. In particular, we calculate rigid-flexible values, i.e. the minimum $a$ such that for $E(1,a')$ with $a'>a$, the embedding problem is determined only by volume. For $1

One of the grand challenges of Mathematics instruction is to provide students with problems that are both accessible and have a reasonably elegant solution. Instructors commonly resort to resources like course textbooks, online-learning platforms, or other automated problem-generating software to select problems for exams and assignments. However, reliance on such tools may result in limited control over problem parameters, potentially yielding intricate solutions that impede students' understanding. This article centers on Linear Algebra, wherein we devise algorithms for reverse engineering matrices of integers with integer outcomes through operations such as the inverse, LU decomposition, and QR decomposition. The focus is on empowering instructors to manipulate matrix properties deliberately, ensuring the creation of problems that enrich instruction and foster student confidence. The intellectual endeavor of reverse engineering such problems, grounded in both theory and matrix properties, proves mutually beneficial for both students and instructors alike.

While alignment algorithms are now commonly used to tune pre-trained language models towards a user's preferences, we lack explanations for the underlying mechanisms in which models become ``aligned'', thus making it difficult to explain phenomena like jailbreaks. In this work we study a popular algorithm, direct preference optimization (DPO), and the mechanisms by which it reduces toxicity. Namely, we first study how toxicity is represented and elicited in a pre-trained language model, GPT2-medium. We then apply DPO with a carefully crafted pairwise dataset to reduce toxicity. We examine how the resulting model averts toxic outputs, and find that capabilities learned from pre-training are not removed, but rather bypassed. We use this insight to demonstrate a simple method to un-align the model, reverting it back to its toxic behavior.

Practitioners commonly align large language models using pairwise preferences, i.e., given labels of the type response A is preferred to response B for a given input. Perhaps less commonly, methods have also been developed for binary feedback, i.e. training models given labels of type response A is good or bad. We show how an existing performant binary feedback method, the Cringe Loss (Adolphs et al., 2022), can be generalized to the pairwise preference setting using a simple soft margin extension. Pairwise Cringe Loss is straightforward to implement and efficient to train, and we find it outperforms state-of-the-art preference optimization algorithms such as PPO and DPO on the AlpacaFarm benchmark. We show that iterations of training of our model are important for improved results, and that we can generalize DPO to Iterative DPO in the same way.

This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of the Gemini family in cross-modal reasoning and language understanding will enable a wide variety of use cases. We discuss our approach toward post-training and deploying Gemini models responsibly to users through services including Gemini, Gemini Advanced, Google AI Studio, and Cloud Vertex AI.

This manuscript aims to formalize and conclude the discussions initiated during the PriTEM workshop 22-23 March 2023. We present important ideas and discussion topics in the context of transactive energy systems. Moreover, the conclusions from the discussions articulate potential aspects to be explored in future studies on transactive energy management. Particularly, these conclusions cover research topics in energy technology and energy informatics, energy law, data law, energy market and socio-psychology that are relevant to the seamless integration of renewable energy resources and the transactive energy systems-in smart microgrids-focusing on distributed frameworks such as peer-to-peer (P2P) energy trading. We clarify issues, identify barriers, and suggest possible solutions to open questions in diversified topics, such as block-chain interoperability, consumer privacy and data sharing, and participation incentivization. Furthermore, we also elaborate challenges associated with cross-disciplinary collaboration and coordination for transactive energy systems, and enumerate the lessons learned from our work so far.

The widespread application of metal additive manufacturing (AM) is limited by the ability to control the complex interactions between the energy source and the feedstock material. Here we develop a generalizable process to introduce nanoscale grooves to the surface of metal powders which increases the powder absorptivity by up to 70% during laser powder bed fusion. Absorptivity enhancements in copper, copper-silver, and tungsten enables energy efficient manufacturing, with printing of pure copper at relative densities up to 92% using laser energy densities as low as 82 J/mm^3. Simulations show the enhanced powder absorptivity results from plasmon-enabled light concentration in nanoscale grooves combined with multiple scattering events. The approach taken here demonstrates a general method to enhance the absorptivity and printability of reflective and refractory metal powders by changing the surface morphology of the feedstock without altering its composition.

Although the convergence of high-performance computing, automation, and machine learning has significantly altered the materials design timeline, transformative advances in functional materials and acceleration of their design will require addressing the deficiencies that currently exist in materials informatics, particularly a lack of standardized experimental data management. The challenges associated with experimental data management are especially true for combinatorial materials science, where advancements in automation of experimental workflows have produced datasets that are often too large and too complex for human reasoning. The data management challenge is further compounded by the multi-modal and multi-institutional nature of these datasets, as they tend to be distributed across multiple institutions and can vary substantially in format, size, and content. To adequately map a materials design space from such datasets, an ideal materials data infrastructure would contain data and metadata describing i) synthesis and processing conditions, ii) characterization results, and iii) property and performance measurements. Here, we present a case study for the low-barrier development of such a dashboard that enables standardized organization, analysis, and visualization of a large data lake consisting of combinatorial datasets of synthesis and processing conditions, X-ray diffraction patterns, and materials property measurements generated at several different institutions. While this dashboard was developed specifically for data-driven thermoelectric materials discovery, we envision the adaptation of this prototype to other materials applications, and, more ambitiously, future integration into an all-encompassing materials data management infrastructure.

This work studies a central extremal graph theory problem inspired by a 1975 conjecture of Erd\H{o}s, which aims to find graphs with a given size (number of nodes) that maximize the number of edges without having 3- or 4-cycles. We formulate this problem as a sequential decision-making problem and compare AlphaZero, a neural network-guided tree search, with tabu search, a heuristic local search method. Using either method, by introducing a curriculum -- jump-starting the search for larger graphs using good graphs found at smaller sizes -- we improve the state-of-the-art lower bounds for several sizes. We also propose a flexible graph-generation environment and a permutation-invariant network architecture for learning to search in the space of graphs., Comment: To appear in the proceedings of IJCAI 2024. First three authors contributed equally, last two authors made equal senior contribution

Many real-world datasets live on high-dimensional Stiefel and Grassmannian manifolds, $V_k(\mathbb{R}^N)$ and $Gr(k, \mathbb{R}^N)$ respectively, and benefit from projection onto lower-dimensional Stiefel (respectively, Grassmannian) manifolds. In this work, we propose an algorithm called Principal Stiefel Coordinates (PSC) to reduce data dimensionality from $ V_k(\mathbb{R}^N)$ to $V_k(\mathbb{R}^n)$ in an $O(k)$-equivariant manner ($k \leq n \ll N$). We begin by observing that each element $\alpha \in V_n(\mathbb{R}^N)$ defines an isometric embedding of $V_k(\mathbb{R}^n)$ into $V_k(\mathbb{R}^N)$. Next, we optimize for such an embedding map that minimizes data fit error by warm-starting with the output of principal component analysis (PCA) and applying gradient descent. Then, we define a continuous and $O(k)$-equivariant map $\pi_\alpha$ that acts as a ``closest point operator'' to project the data onto the image of $V_k(\mathbb{R}^n)$ in $V_k(\mathbb{R}^N)$ under the embedding determined by $\alpha$, while minimizing distortion. Because this dimensionality reduction is $O(k)$-equivariant, these results extend to Grassmannian manifolds as well. Lastly, we show that the PCA output globally minimizes projection error in a noiseless setting, but that our algorithm achieves a meaningfully different and improved outcome when the data does not lie exactly on the image of a linearly embedded lower-dimensional Stiefel manifold as above. Multiple numerical experiments using synthetic and real-world data are performed., Comment: 26 pages, 8 figures, comments welcome!

How do sequence models represent their decision-making process? Prior work suggests that Othello-playing neural network learned nonlinear models of the board state (Li et al., 2023). In this work, we provide evidence of a closely related linear representation of the board. In particular, we show that probing for "my colour" vs. "opponent's colour" may be a simple yet powerful way to interpret the model's internal state. This precise understanding of the internal representations allows us to control the model's behaviour with simple vector arithmetic. Linear representations enable significant interpretability progress, which we demonstrate with further exploration of how the world model is computed.

The rapid accessibility of portable and affordable retinal imaging devices has made early differential diagnosis easier. For example, color funduscopy imaging is readily available in remote villages, which can help to identify diseases like age-related macular degeneration (AMD), glaucoma, or pathological myopia (PM). On the other hand, astronauts at the International Space Station utilize this camera for identifying spaceflight-associated neuro-ocular syndrome (SANS). However, due to the unavailability of experts in these locations, the data has to be transferred to an urban healthcare facility (AMD and glaucoma) or a terrestrial station (e.g, SANS) for more precise disease identification. Moreover, due to low bandwidth limits, the imaging data has to be compressed for transfer between these two places. Different super-resolution algorithms have been proposed throughout the years to address this. Furthermore, with the advent of deep learning, the field has advanced so much that x2 and x4 compressed images can be decompressed to their original form without losing spatial information. In this paper, we introduce a novel model called Swin-FSR that utilizes Swin Transformer with spatial and depth-wise attention for fundus image super-resolution. Our architecture achieves Peak signal-to-noise-ratio (PSNR) of 47.89, 49.00 and 45.32 on three public datasets, namely iChallenge-AMD, iChallenge-PM, and G1020. Additionally, we tested the model's effectiveness on a privately held dataset for SANS provided by NASA and achieved comparable results against previous architectures., Comment: Accepted in 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023

Recent progress in large language models (LLMs) has enabled the deployment of many generative NLP applications. At the same time, it has also led to a misleading public discourse that ``it's all been solved.'' Not surprisingly, this has, in turn, made many NLP researchers -- especially those at the beginning of their careers -- worry about what NLP research area they should focus on. Has it all been solved, or what remaining questions can we work on regardless of LLMs? To address this question, this paper compiles NLP research directions rich for exploration. We identify fourteen different research areas encompassing 45 research directions that require new research and are not directly solvable by LLMs. While we identify many research areas, many others exist; we do not cover areas currently addressed by LLMs, but where LLMs lag behind in performance or those focused on LLM development. We welcome suggestions for other research directions to include: https://bit.ly/nlp-era-llm, Comment: Accepted at COLING 2024