Zero-Shot Learning for IMU-Based Activity Recognition Using Video Embeddings

The Activity Recognition Chain generally precludes the challenging scenario of recognizing new activities that were unseen during training, despite this scenario being a practical and common one as users perform diverse activities at test time. A few prior works have adopted zero-shot learning methods for IMU-based activity recognition, which work by relating seen and unseen classes through an auxiliary semantic space. However, these methods usually rely heavily on a hand-crafted attribute space which is costly to define, or a learnt semantic space based on word embedding, which lacks motion-related information crucial for distinguishing IMU features. Instead, we propose a strategy to exploit videos of human activities to construct an informative semantic space. With our approach, knowledge from state-of-the-art video action recognition models is encoded into video embeddings to relate seen and unseen activity classes. Experiments on three public datasets find that our approach outperforms other learnt semantic spaces, with an additional desirable feature of scalability, as recognition performance is seen to scale with the amount of data used. More generally, our results indicate that exploiting information from the video domain for IMU-based tasks is a promising direction, with tangible returns in a zero-shot learning scenario.

Universality of Riemann solutions in porous media

Universality, a desirable feature in any system. For decades, elusive measurements of three-phase flows have yielded countless permeability models that describe them. However, the equations governing the solution of water and gas co-injection has a robust structure. This universal structure stands for Riemann problems in green oil reservoirs. In the past we established a large class of three phase flow models including convex Corey permeability, Stone I and Brooks–Corey models. These models share the property that characteristic speeds become equal at a state somewhere in the interior of the saturation triangle. Here we construct a three-phase flow model with unequal characteristic speeds in the interior of the saturation triangle, equality occurring only at a point of the boundary of the saturation triangle. Yet the solution for this model still displays the same universal structure, which favors the two possible embedded two-phase flows of water-oil or gas-oil. We focus on showing this structure under the minimum conditions that a permeability model must meet. This finding is a guide to seeking a purely three-phase flow solution maximizing oil recovery.

Modeling and evaluating beat gestures for social robots

Natural gestures are a desirable feature for a humanoid robot, as they are presumed to elicit a more comfortable interaction in people. With this aim in mind, we present in this paper a system to develop a natural talking gesture generation behavior. A Generative Adversarial Network (GAN) produces novel beat gestures from the data captured from recordings of human talking. The data is obtained without the need for any kind of wearable, as a motion capture system properly estimates the position of the limbs/joints involved in human expressive talking behavior. After testing in a Pepper robot, it is shown that the system is able to generate natural gestures during large talking periods without becoming repetitive. This approach is computationally more demanding than previous work, therefore a comparison is made in order to evaluate the improvements. This comparison is made by calculating some common measures about the end effectors’ trajectories (jerk and path lengths) and complemented by the Fréchet Gesture Distance (FGD) that aims to measure the fidelity of the generated gestures with respect to the provided ones. Results show that the described system is able to learn natural gestures just by observation and improves the one developed with a simpler motion capture system. The quantitative results are sustained by questionnaire based human evaluation.

Utilization of Eight-Variable Karnaugh Maps in the Exploration of Problems of Qualitative Comparative Analysis

Qualitative Comparative Analysis (QCA) is an emergent methodology of diverse applications in many disciplines. However, its premises and techniques are continuously subject to discussion, debate, and (even) dispute. We use a regular and modular Karnaugh map to explore a prominent recently-posed eight-variable QCA problem. This problem involves a partially-defined Boolean function (PDBF), that is dominantly unspecified. Without using the algorithmic integer-programming approach, we devise a simple heuristic map procedure to discover minimal sets of supporting variables. The eight-variable problem studied herein is shown to have at least two distinct such sets, with cardinalities of 4 and 3, respectively. For these two sets, the pertinent function is still a partially-defined Boolean function (PDBF), equivalent to 210 = 1024 completely-specified Boolean functions (CSBFs) in the first case, and to four CSBFs only in the second case. We obtained formulas for the four functions of the second case, and a formula for a sample fifth function in the first case. Although only this fifth function is unate, each of the five functions studied does not have any non-essential prime implicant, and hence each of them enjoys the desirable feature of having a single IDF that is both a unique minimal sum and the complete sum. According to our scheme of first identifying a minimal set of supporting variables, we avoided the task of drawing prime-implicant loops on the initial eight-variable map, and postponed this task till the map became dramatically reduced in size. Our map techniques and results are hopefully of significant utility in future QCA applications.

Agile Tasking of Robotic Systems with Explicit Autonomy

Task agility is an increasingly desirable feature for robots in application domains such as manufacturing. The Canonical Robot Command Language (CRCL) is a lightweight information model built for agile tasking of robotic systems. CRCL replaces the underlying complex proprietary robot programming interface with a standard interface. In this paper, we exchange the automated planning component that CRCL used in the past for a rational agent in the Gwendolen agent programming language, thus providing greater possibilities for formal verification and explicit autonomy. We evaluate our approach by performing agile tasking in a kitting case study.

Mutations in hik26 and slr1916 lead to high-light stress tolerance in Synechocystis sp. PCC6803

Increased tolerance to light stress in cyanobacteria is a desirable feature for their applications. Here, we obtained a high light tolerant (Tol) strain of Synechocystis sp. PCC6803 through an adaptive laboratory evolution, in which the cells were repeatedly sub-cultured for 52 days under high light stress conditions (7000 to 9000 μmol m−2 s−1). Although the growth of the parental strain almost stopped when exposed to 9000 μmol m−2 s−1, no growth inhibition was observed in the Tol strain. Excitation-energy flow was affected because of photosystem II damage in the parental strain under high light conditions, whereas the damage was alleviated and normal energy flow was maintained in the Tol strain. The transcriptome data indicated an increase in isiA expression in the Tol strain under high light conditions. Whole genome sequence analysis and reverse engineering revealed two mutations in hik26 and slr1916 involved in high light stress tolerance in the Tol strain.

Flux-Modulated Permanent Magnet Machines: Challenges and Opportunities

High torque density is a desirable feature of electrical machines used in traction applications, such as electric vehicle (EV)/hybrid electric vehicle (HEV) propulsion, wind turbines, more electric aircrafts, etc. The flux-modulated permanent magnet (FMPM) machine is considered as one of the most promising candidates to achieve high torque density. The incorporated gearing effect is ideal in reducing the rotating speed and amplifying the output torque of the FMPM machines. This paper aims at a comprehensive review of the topology evolution of the FMPM machines. Based on different structures, the FMPM machines are grouped into four categories: surface-type FMPM machines, spoke-type FMPM machines, partitioned stator FMPM machines, and bidirectional FMPM machines. The operating principles, advantages, drawbacks, and major applications of the FMPM machines are discussed in detail. In the end, the current state of the art, opportunities, challenges, and future trends of the FMPM machines are discussed. Therefore, this paper offers a systematic guidance on the selection and design of the FMPM machines.

Supramolecular tuning of Thioflavin-T aggregation hosted by polystyrene sulfonate

A tunable and controllable emission is an extremely desirable feature for advanced functional materials that finds usage in optoelectronic utilization, fluorescence probing/ sensing, drug-delivery monitoring, etc. In the present contribution,...

New emotional model environment for navigation in a virtual reality

Emotions are commonly considered to be the most expressive of everyday human experiences, giving a sense of fullness and reality of life. The ability to recognize human emotions as a manifestation of higher intelligence is desirable feature. There are several models of emotional experience that can become the basis for building a universal emotional recognition system. In this article, we check the correctness of the designed emotional model. We check the evaluation of the system’s operation by human observers.