We present a scalable, end-to-end workflow for protein design. By augmenting protein sequences with natural language descriptions of their biochemical properties, we train generative models that can be preferentially aligned with protein fitness landscapes. Through complex experimental-and simulation-based observations, we integrate these measures as preferred parameters for generating new protein variants and demonstrate our workflow on five diverse supercomputers. We achieve >1 ExaFLOPS sustained performance in mixed precision on each supercomputer and a maximum sustained performance of 4.11 Ex-aFLOPS and peak performance of 5.57 ExaFLOPS. We establish the scientific performance of our model on two tasks: (1) across a predetermined benchmark dataset of deep mutational scanning experiments to optimize the fitness-determining mutations in the yeast protein HIS7, and (2) in optimizing the design of the enzyme malate dehydrogenase to achieve lower activation barriers (and therefore increased catalytic rates) using simulation data. Our implementation thus sets high watermarks for multimodal protein design workflows.

Machine-learning (ML) hardware and software system demand is burgeoning. Driven by ML applications, the number of different ML inference systems has exploded. Over 100 organizations are building ML inference chips, and the systems that incorporate existing models span at least three orders of magnitude in power consumption and five orders of magnitude in performance; they range from embedded devices to data-center solutions. Fueling the hardware are a dozen or more software frameworks and libraries. The myriad combinations of ML hardware and ML software make assessing ML-system performance in an architecture-neutral, representative, and reproducible manner challenging. There is a clear need for industry-wide standard ML benchmarking and evaluation criteria. MLPerf Inference answers that call. In this paper, we present our benchmarking method for evaluating ML inference systems. Driven by more than 30 organizations as well as more than 200 ML engineers and practitioners, MLPerf prescribes a set of rules and best practices to ensure comparability across systems with wildly differing architectures. The first call for submissions garnered more than 600 reproducible inference-performance measurements from 14 organizations, representing over 30 systems that showcase a wide range of capabilities. The submissions attest to the benchmark’s flexibility and adaptability.

This paper proposes interactive decoration design support system introducing interactive evolutionary computation. Evaluation problem existing human can be evaluated synthesizing multi objectives. The proposed system made it importance dividing the human evaluation into two part, which is including regular level of design quality and user's subjective evaluation by affective or Kansei image, and assumed to acquire good results by executing in serial order in a certain evaluation phase. From the experimental results, effectiveness of proposed methodology involving evolution by using fitness evaluation of decoration designs for quality and evolution by user evaluation are confirmed.

We have developed a rough set model for analyzing an information system in which some conditions as well as decision values, are missing. Current studies have focused mainly on the missing of condition data but seem to ignore the missing of decision data. The common approach is to remove objects with no decision values because such objects are apparently considered fruitless from the decision-making standpoint. However, this deletion may lead to the risk of information loss. We observe that such a situation is somewhat similar to the semi-supervised situation in the sense that some objects are characterized by complete decision data while some are not. Considering both kinds of objects from a probabilistic view, we predict potential candidates for missing values by comparing measurements of two factors, local decision belief and universal decision belief, with a parameter threshold α. These possible decision candidates help to form a relative dissimilarity relation, which measures the unlikeness of pairs of objects rather than their likeness. Contrasting with the other approaches, rough set definitions based on this relation do not approximate the target set but its complement instead. The knowledge acquisition induced by the common approach and the proposed approach is compared, and the result shows that the latter can overcome some limitations of the former. This approach is new and flexible to deal with missing decision information.

The original rough set theory deals with precise and complete data, while real applications frequently contain imperfect information. A typical imperfect data studied in rough set research is the missing values. Though there are many ideas proposed to solve the issue in the literature, the paper adopts a probabilistic approach, because it can incorporate other types of imperfect data including imprecise and uncertain values in a single approach. The paper first discusses probabilities of attribute values assuming different type of attributes in real applications, and proposes a generalized method of probability of matching. It also discusses the case of continuous data as well as discrete one. The proposed probability of matching could be used for defining valued tolerance/similarity relations in rough set approaches.

In this paper, a descriptive decision-making model under uncertainty is proposed which incorporates two types of decision attitudes for uncertainty; one is an attitude about ignorance (optimism/pessimism) and the other one is about risk (risk-seeking and risk-aversion). At first, Evidential Decision Making Problem (EDMP) has been defined where Dempster-Shafer Theory (DST) has been used to represent uncertainty. Then probability approximation approach of solving EDMP is shown. For deciding the decision weights in different attitudes of decision maker, Ordered Weighted Averaging (OWA) operator has been used. Later on, Prospect Theory has been applied to accomplish a descriptive decision-making model. To show the effectiveness of our approach, a real life decision problem of travelers' route choice from a set of alternatives has also been provided.

Unconstrained growth of synaptic connectivity and the lack of references to synaptic depression in Hebb's postulate has diminished its value as a learning algorithm. While spike timing dependent plasticity and other synaptic scaling mechanisms have been studying the possibility of regulating synaptic activity on neuronal level, we studied the possibility of regulating the synaptic activity of Hebb's neurons on dynamic stochastic computational synapses. The study was conducted on fully connected network with four artificial neurons where each neuron consisted of thousands of artificial stochastic synapses that are modeled with transmitters and receptors. The synapses updated their stochastic states dynamically according to the spike arrival time to that synapses. The activity of these synapses was regulated by a new stability promoting mechanism. Results support the following findings: (i) the synchronous activity between presynaptic (cell A) and postsynaptic (cell B) neuron increases the activity of A. (ii) Asynchronous activation of these two neurons decreases A's activity if one of the following conditions are satisfied (a). if activity of the other presynaptic neurons of the postsynaptic neuron B is asynchronous with the A's activity or (b) if B is in a depressed state when activity of presynaptic neuron A is increased. (iii) the introduced stability promoting mechanism exhibited similar to the Homeostatic synaptic plasticity process and encouraged the emergence of Hebb's postulate and its anti-Hebbian mechanisms. Further, we demonstrated the metabolic changes that could occur inside Hebb's neurons when such an activity takes place on a dynamic stochastic neural network.

The paper discusses an idea of representing brand image on a computer and simulating associations and interactions among multiple pieces of brand image. Brand image is represented using a fuzzy set based on the theory of brand personality, which is a theory to represent brand image indirectly by a set of human characteristics associated with a brand. An convenient feature of the representation is generality that image of any kind of brands could be defined on the same universal set. The interactions among multiple pieces of image are simulated using the framework of Conceptual Fuzzy Set which is realized as combination of two Fuzzy Bidirectional Associative Memories.

The Multiple Instruction Stream Processor (MISP) architecture introduces the sequencer as a new class of architectural resource, and provides a minimalist user-level MIMD instruction set extension for application programs to directly control execution of concurrent instruction streams on these sequencers. As with classic architectural resources, namely, registers and memory, the sequencer architectural resource can be subject to virtualization. This paper details the idea of Sequencer Virtualization (SV), a foundational architectural support to decouple architectural virtual sequencers from physical sequencers. SV enables more efficient utilization of sequencer resources at the microarchitectural level while maintaining a consistent programming interface at the architectural level. To evaluate the key tradeoffs for SV, we conduct extensive experiments by implementing a prototype SV system using a custom firmware on a large-scale multiprocessor system. Using the prototype SV system, we demonstrate that SV improves efficiency in sequencer utilization while incurring little performance overhead. In particular, for a set of real multithreaded workloads, SV can significantly improve sequencer utilization, achieving an average of 32% better wall-clock performance than MISP without SV support in a multi-programming environment.

Querying in heterogeneous sensor networks is a challenging research issue due to a variety of real-world queries depending on users' preferences. Examples of queries are weather, nearby restaurants, navigation, etc. Users may ask for a breezy path starting from distinct points to distinct train stations. Sensing data collected from wide areas (city or country level) are needed to provide the real-world search as a service for users. Therefore, another challenge is a collaboration of heterogenous sensor networks because deploying sensors in wide areas is impractical. SensorMap [1] collects sensing data from independent sensor networks by allowing data owners to publish their data on the web. Sensing data from all owners are sent to a centric storage in SensorMap. However, distributed data storage is likely to be an efficient solution comparing to a centric storage in which bottleneck always occurs.

This paper concerns a control strategy for multiple systems that affect the condition of the air in a room such as electric heaters and floor heating systems. Based on a resident lifestyle and case-based approach, the control system with the proposed strategy optimizes the weighted summation of three evaluation indices: cost, the time needed to reach a comfortable state and the rate of the comfortable state where the comfortable state is defined by well-known PMV (predicted mean vote). This paper discusses how the three indices are evaluated and describes the roles of various components in the proposed system. This paper also presents the user interface for expressing a user's lifestyle and it touches on future issues.

This paper deals with the fast and robust detection of the traffic sign images. A new technique called geometric fragmentation is proposed to detect the red circular traffic signs. It detects the outer ellipses of the signs by combining the left and right fragments of the ellipse objects. A search based on the geometric fragmentation is used to find the ellipse fragments. This search is somewhat similar to genetic algorithm (GA) in the sense that it employs the terms of individual, population, crossover, and objective function usually used in GA. To increase the accuracy and reduce the computational time, a new objective function is introduced for evaluating the individuals. The algorithm was tested for detecting the red circular traffic signs from the real scene image. The experimental results show that the proposed algorithm has a higher detection rate with a lower computational cost compared with the referential genetic algorithm-based ellipse detection

Human interface design for household appliances is becoming far more difficult than the one in the past, because they have so many functions and the control panels are so compact these days. The user cannot remember all the functions the appliances have, as well as enough buttons, switches and/or displays cannot be placed on the small control panels. The authors have proposed a novel human interface paradigm named push like talking. PLT has buttons each of which has a word expressing a concept important to operate the appliance, and the user pushes the buttons with a word not to select a function, but to talk to the appliance. PLT reasons the intention of the user from the words on the pushed buttons. The paper proposes a possibilistic model representing the relations between the intentions and the words, and discusses a way to reason the intention from the words.

Many events in the world occur with some quantity that shows a level of the occurrence. The paper discusses reasoning with the normalized level of the occurrence, which we call intensity, and proposes intensity reasoning by constraint propagation based on causal relationships. The knowledge used in the reasoning is given by a directed acyclic graph called causal network. Each node in the network expresses an event with an intensity in [0,1], and each are between two nodes does a causal relation defined by a function that gives the relation between intensities of those nodes. The reasoning is conducted by constraint propagation to derive a range of intensity of an arbitrarily chosen node when those of some other nodes are given.