Test-time adaptation aims to improve model robustness under distribution shifts by adapting models with access to unlabeled target samples. A primary cause of performance degradation under such shifts is the model's reliance on features that lack a direct causal relationship with the prediction target. We introduce Test-time Adaptation by Causal Trimming (TACT), a method that identifies and removes non-causal components from representations for test distributions. TACT applies data augmentations that preserve causal features while varying non-causal ones. By analyzing the changes in the representations using Principal Component Analysis, TACT identifies the highest variance directions associated with non-causal features. It trims the representations by removing their projections on the identified directions, and uses the trimmed representations for the predictions.

Large Language Models (LLMs) often do not perform well on queries that require the aggregation of information across texts. To better evaluate this setting and facilitate modeling efforts, we introduce TACT - Text And Calculations through Tables, a dataset crafted to evaluate LLMs' reasoning and computational abilities using complex instructions. TACT contains challenging instructions that demand stitching information scattered across one or more texts, and performing complex integration on this information to generate the answer. We construct this dataset by leveraging an existing dataset of texts and their associated tables. For each such tables, we formulate new queries, and gather their respective answers.

Tactile sensing is fundamental to robotic systems, enabling interactions through physical contact in multiple tasks. Despite its importance, achieving high-resolution, large-area tactile sensing remains challenging. Electrical Impedance Tomography (EIT) has emerged as a promising approach for large-area, distributed tactile sensing with minimal electrode requirements which can lend itself to addressing complex contact problems in robotics. However, existing EIT-based tactile reconstruction methods often suffer from high computational costs or depend on extensive annotated simulation datasets, hindering its viability in real-world settings. To address this shortcoming, here we propose a Pre-trained Transformer for EIT-based Tactile Reconstruction (PTET), a learning-based framework that bridges the simulation-to-reality gap by leveraging self-supervised pretraining on simulation data and fine-tuning with limited real-world data. In simulations, PTET requires 99.44 percent fewer annotated samples than equivalent state-of-the-art approaches (2,500 vs. 450,000 samples) while achieving reconstruction performance improvements of up to 43.57 percent under identical data conditions. Fine-tuning with real-world data further enables PTET to overcome discrepancies between simulated and experimental datasets, achieving superior reconstruction and detail recovery in practical scenarios. The improved reconstruction accuracy, data efficiency, and robustness in real-world tasks establish it as a scalable and practical solution for tactile sensing systems in robotics, especially for object handling and adaptive grasping under varying pressure conditions.

Large Language Models (LLMs) often do not perform well on queries that require the aggregation of information across texts. To better evaluate this setting and facilitate modeling efforts, we introduce TACT - Text And Calculations through Tables, a dataset crafted to evaluate LLMs' reasoning and computational abilities using complex instructions. TACT contains challenging instructions that demand stitching information scattered across one or more texts, and performing complex integration on this information to generate the answer. We construct this dataset by leveraging an existing dataset of texts and their associated tables. For each such tables, we formulate new queries, and gather their respective answers.

Model merging aims to efficiently combine the weights of multiple expert models, each trained on a specific task, into a single multi-task model, with strong performance across all tasks. When applied to all but the last layer of weights, existing methods -- such as Task Arithmetic, TIES-merging, and TALL mask merging -- work well to combine expert models obtained by fine-tuning a common foundation model, operating within a "local" neighborhood of the foundation model. This work explores the more challenging scenario of "non-local" merging, which we find arises when an expert model changes significantly during pretraining or where the expert models do not even share a common foundation model. We observe that standard merging techniques often fail to generalize effectively in this non-local setting, even when accounting for permutation symmetries using standard techniques. We identify that this failure is, in part, due to "variance collapse", a phenomenon identified also in the setting of linear mode connectivity by Jordan et al. (2023). To address this, we propose a multi-task technique to re-scale and shift the output activations of the merged model for each task, aligning its output statistics with those of the corresponding task-specific expert models. Our experiments demonstrate that this correction significantly improves the performance of various model merging approaches in non-local settings, providing a strong baseline for future research on this problem.

Large Language Models (LLMs) often do not perform well on queries that require the aggregation of information across texts. To better evaluate this setting and facilitate modeling efforts, we introduce TACT - Text And Calculations through Tables, a dataset crafted to evaluate LLMs' reasoning and computational abilities using complex instructions. TACT contains challenging instructions that demand stitching information scattered across one or more texts, and performing complex integration on this information to generate the answer. We construct this dataset by leveraging an existing dataset of texts and their associated tables. For each such tables, we formulate new queries, and gather their respective answers. We demonstrate that all contemporary LLMs perform poorly on this dataset, achieving an accuracy below 38\%. To pinpoint the difficulties and thoroughly dissect the problem, we analyze model performance across three components: table-generation, Pandas command-generation, and execution. Unexpectedly, we discover that each component presents substantial challenges for current LLMs. These insights lead us to propose a focused modeling framework, which we refer to as IE as a tool. Specifically, we propose to add "tools" for each of the above steps, and implement each such tool with few-shot prompting. This approach shows an improvement over existing prompting techniques, offering a promising direction for enhancing model capabilities in these tasks.

The 3D modelling of indoor environments and the generation of process simulations play an important role in factory and assembly planning. In brownfield planning cases existing data are often outdated and incomplete especially for older plants, which were mostly planned in 2D. Thus, current environment models cannot be generated directly on the basis of existing data and a holistic approach on how to build such a factory model in a highly automated fashion is mostly non-existent. Major steps in generating an environment model in a production plant include data collection and pre-processing, object identification as well as pose estimation. In this work, we elaborate a methodical workflow, which starts with the digitalization of large-scale indoor environments and ends with the generation of a static environment or simulation model. The object identification step is realized using a Bayesian neural network capable of point cloud segmentation. We elaborate how the information on network uncertainty generated by a Bayesian segmentation framework can be used in order to build up a more accurate environment model. The steps of data collection and point cloud segmentation as well as the resulting model accuracy are evaluated on a real-world data set collected at the assembly line of a large-scale automotive production plant. The segmentation network is further evaluated on the publicly available Stanford Large-Scale 3D Indoor Spaces data set. The Bayesian segmentation network clearly surpasses the performance of the frequentist baseline and allows us to increase the accuracy of the model placement in a simulation scene considerably.

Forbes' Jillian D'Onfro leads a panel of AI industry experts (left to right) Marco Casalaina, Salesforce, Chuck Ganapathi, Tact.ai, and Lorrissa Horton, Cisco. Today's voice-powered AI assistant has many names--Siri, Alexa, Cortana--but as this developing technology becomes ubiquitous in both consumer and enterprise environments, Chuck Ganapathi has a suggestion for his industry colleagues: "Let's not pretend it's a human," the founder and CEO of Tact said Monday during the Voice AI in the Enterprise panel at the Forbes CIO Summit in Half Moon Bay, California, taking a jab at Google's eerily lifelike Duplex AI system. What's more, he said, the enterprise must be careful not to hark back to the secretary pools of old: Voice assistants at work shouldn't automatically sound like they're women. His company intentionally gave its AI-based customer-relationship management system a gender-free name and gives users multiple voice options at setup. He also commended the work of a European agency called Virtue that launched the first "genderless" digital assistant voice this spring.

It's not every day that large competitors become investors in the same company. This week, however, one CRM company attracted funding in a Series C round from four of the biggest: Microsoft, Amazon, Salesforce and Comcast. The venture capital arms of those tech giants - as well as a few other well-known VC investors like Accel Partners and Redpoint Ventures, plowed $27 million into Tact.ai, What's so special about Tact.ai? The company's technology uses artificial intelligence to solve a CRM headache that many of my clients consistently complain about: getting their sales people to use the system!