Добірка наукової літератури з теми "General-purpose agent"

Abstract The use of the sol—gel process on general-purpose grade rubbers is reviewed in the absence or presence of silane coupling agents. The sol—gel reactions of tetraethoxysilane (TEOS) in epoxidized natural rubber (ENR), styrene—butadiene rubber (SBR) or butadiene rubber (BR) vulcanizates produced silica generated in situ. This silica was found to be a good reinforcing agent by investigating tensile and dynamic mechanical properties and morphology observation by transmission electron microscopy (TEM). The amount of silica formed was limited by the degree of swelling of the rubber vulcanizate by TEOS which was the precursor of the silica. However, the dispersion of silica generated in situ was better than conventionally added silica due to its formation in place. Also, it was noted that the diameter distribution of in situ silica was monodispersed. Silane coupling agents, such as mercaptosilane, aminosilane, and bis(3-triethoxysilylpropyl) tetrasulfide, were compounded in the vulcanizates and their effects on silica generated in situ were evaluated. Their effects were significant. The dispersion of the silica in the rubbery matrix became better and the particle size became smaller and monodispersed, as observed by TEM, which improved mechanical properties. The superior properties of silica generated in situ have been studied further to elucidate the mechanism of reinforcement.

The purpose of the study is to study the possibilities of multigenerational optimization of behavior control systems for agents of general artificial intelligence capable of independently solving a universal range of tasks in a real environment. The main principles of ontophylogenetic synthesis of control systems for agents of general artificial intelligence based on multi-agent neurocognitive architectures have been developed. Methods and algorithms for synthesizing the phenotypes of control systems of intelligent agents according to their genotypes are proposed. A software package for simulating the processes of ontophylogenetic synthesis of multi-agent neurocognitive architectures has been developed and experiments have been carried out to create phenotypes of intelligent agents based on them. A complex genome of an intelligent agent has been developed, the features of a multichromosome genetic algorithm for organizing calculations in the paradigm of multigenerational optimization of multiagent neurocognitive architectures have been established and substantiated. It is shown that multigenerational optimization of the multi-agent neurocognitive architecture of intelligent agents can contribute to the achievement of adaptive resistance to the operating conditions of a general artificial intelligence agent, provide the synthesis of its suboptimal structural and functional scheme, accelerate learning and algorithms for finding solutions to a universal range of problems solved by this agent in its ecological niche.

AbstractOne of the essential aspect in biological agents is dynamic stability. This aspect, called homeostasis, is widely discussed in ethology, neuroscience and during the early stages of artificial intelligence. Ashby’s homeostats are general-purpose learning machines for stabilizing essential variables of the agent in the face of general environments. However, despite their generality, the original homeostats couldn’t be scaled because they searched their parameters randomly. In this paper, first we re-define the objective of homeostats as the maximization of a multi-step survival probability from the view point of sequential decision theory and probabilistic theory. Then we show that this optimization problem can be treated by using reinforcement learning algorithms with special agent architectures and theoretically-derived intrinsic reward functions. Finally we empirically demonstrate that agents with our architecture automatically learn to survive in a given environment, including environments with visual stimuli. Our survival agents can learn to eat food, avoid poison and stabilize essential variables through theoretically-derived single intrinsic reward formulations.

Because of the poor reconfigurability of the existing test resource models and the faultiness of the intelligent modeling methods in automatic test system, the modeling method based on Multi-Agent was studied. By analyzing the characters of test resources, the test resource models were classified as device model, adapter model and configuration model. The typical test resource model based on Multi-Agent with three layers and five agents was built. The structure and function of system management agent, model reconfiguration agent, function management agent, channel control agent and function realization agent in the model were analyzed. Then its application in Intelligent Virtual Instrument was introduced. This model based on Multi-Agent has good reconfigurability and intelligence, and it can be used to build the general-purpose and intelligent automatic test system.

This paper presents a novel implementation of an embedded Java microarchitecture for secure, realtime, and multi-threaded applications. Together with the support of modern features of object-oriented languages, such as exception handling, automatic garbage collection and interface types, a general-purpose platform is established which also fits for the agent concept. Especially, considering real-time issues, new techniques have been implemented in our Java microarchitecture, such as an integrated stack and thread management for fast context switching, concurrent garbage collection for real-time threads and autonomous control flows through preemptive round-robin scheduling.

L’apprentissage par renforcement conventionnel implique l’entraînement d’un agent unimodal sur une tâche unique et bien définie, guidé par un signal de récompense optimisé pour le gradient. Ce cadre ne nous permet pas d’envisager un agent d’apprentissage adapté aux problèmes du monde réel impliquant des flux de diverses modalités, des tâches multiples, souvent mal définies, voire pas définies du tout. C’est pourquoi nous préconisons une transition vers un cadre plus général, visant à créer des algorithmes d’apprentissage par renforcement plus adaptables et intrinsèquement polyvalents. Pour progresser dans cette direction, nous identifions deux domaines d’intérêt principaux. Le premier est l’amélioration de l’exploration, qui permet à l’agent d’apprendre de l’environnement en dépendant le moins possible du signal de récompense. Nous présentons Latent Go-Explore (LGE), une généralisation de l’algorithme Go-Explore qui, malgré ses résultats impressionnants, était limité par une forte contrainte de connaissance du domaine. LGE atténue ces limitations et permet une application plus large dans un cadre plus général. LGE démontre son efficacité et sa polyvalence accrues en surpassant de manière significative les lignes de base dans tous les environnements testés. Le deuxième domaine d’intérêt est celui de la conception d’un agent polyvalent qui peut fonctionner dans une variété d’environnements, impliquant ainsi une structure multimodale et transcendant même le cadre séquentiel conventionnel de l’apprentissage par renforcement. Nous présentons Jack of All Trades (JAT), une architecture multimodale basée Transformers, spécialement conçue pour les tâches de décision séquentielle. En utilisant un seul ensemble de poids, JAT démontre sa robustesse et sa polyvalence, rivalisant avec son unique référence sur plusieurs benchmarks d’apprentissage par renforcement et montrant même des performances prometteuses sur des tâches de vision et textuelles. Nous pensons que ces deux contributions constituent une étape importante vers une approche plus générale de l’apprentissage par renforcement. En outre, nous présentons d’autres avancées méthodologiques et techniques qui sont étroitement liées à notre question de recherche initiale. La première est l’introduction d’un ensemble d’environnements robotiques simulés à récompense éparse, conçus pour fournir à la communauté les outils nécessaires à l’apprentissage dans des conditions de faible supervision. Trois ans après son introduction, cette contribution a été largement adoptée par la communauté et continue de faire l’objet d’une maintenance et d’un support actifs. D’autre part, nous présentons Open RL Benchmark, notre initiative pionnière visant à fournir un ensemble complet et entièrement enregistré d’expériences d’apprentissage par renforcement, allant au-delà des données typiques pour inclure toutes les métriques spécifiques à l’algorithme et au système. Ce benchmark vise à améliorer l’efficacité de la recherche en fournissant des données prêtes à l’emploi et en v vi facilitant la reproductibilité précise des expériences. Grâce à son approche communautaire, il est rapidement devenu une ressource importante, documentant plus de 25 000 exécutions. Ces avancées techniques et méthodologiques, associées aux contributions scientifiques décrites ci-dessus, visent à promouvoir une approche plus générale de l’apprentissage par renforcement et, nous l’espérons, représentent une étape significative vers le développement à terme d’un agent plus opérationnel
Conventional Reinforcement Learning (RL) involves training a unimodal agent on a single, well-defined task, guided by a gradient-optimized reward signal. This framework does not allow us to envisage a learning agent adapted to real-world problems involving diverse modality streams, multiple tasks, often poorly defined, sometimes not defined at all. Hence, we advocate for transitioning towards a more general framework, aiming to create RL algorithms that more inherently versatile.To advance in this direction, we identify two primary areas of focus. The first aspect involves improving exploration, enabling the agent to learn from the environment with reduced dependence on the reward signal. We present Latent Go-Explore (LGE), an extension of the Go-Explore algorithm. While Go-Explore achieved impressive results, it was constrained by domain-specific knowledge. LGE overcomes these limitations, offering wider applicability within a general framework. In various tested environments, LGE consistently outperforms the baselines, showcasing its enhanced effectiveness and versatility. The second focus is to design a general-purpose agent that can operate in a variety of environments, thus involving a multimodal structure and even transcending the conventional sequential framework of RL. We introduce Jack of All Trades (JAT), a multimodal Transformer-based architecture uniquely tailored to sequential decision tasks. Using a single set of weights, JAT demonstrates robustness and versatility, competing its unique baseline on several RL benchmarks and even showing promising performance on vision and textual tasks. We believe that these two contributions are a valuable step towards a more general approach to RL. In addition, we present other methodological and technical advances that are closely related to our core research question. The first is the introduction of a set of sparsely rewarded simulated robotic environments designed to provide the community with the necessary tools for learning under conditions of low supervision. Notably, three years after its introduction, this contribution has been widely adopted by the community and continues to receive active maintenance and support. On the other hand, we present Open RL Benchmark, our pioneering initiative to provide a comprehensive and fully tracked set of RL experiments, going beyond typical data to include all algorithm-specific and system metrics. This benchmark aims to improve research efficiency by providing out-of-the-box RL data and facilitating accurate reproducibility of experiments. With its community-driven approach, it has quickly become an important resource, documenting over 25,000 runs.These technical and methodological advances, along with the scientific contributions described above, are intended to promote a more general approach to Reinforcement Learning and, we hope, represent a meaningful step toward the eventual development of a more operative RL agent

L’auto-organisation entre les cellules permet d’expliquer la morphogenèse des tissus cellulaires, comme le phénomène de l’embryogenèse. Cependant, il n’existe pas de consensus sur la nature des interactions entre les cellules amenant à cette auto-organisation. La modélisation et la simulation in silico offrent un support formel aidant le biologiste dans sa compréhension du phénomène et donnent des arguments en faveur d’une théorie ou d’une autre. La mise en œuvre informatique de processus biologiques permet, en retour, d’améliorer les modèles informatiques existants. Les systèmes multi-agents sont des modèles informatiques qui représentent chaque entité (agent) du système de manière explicite. Les agents sont exécutés de manière autonome et en interaction avec les autres. La mise en œuvre d’un système multi-agents permet de simuler des agents cellules où les interactions reposent sur la consommation et la production de molécules, mais aussi sur l’adhésion et la pression que les cellules exercent les unes sur les autres. Un agent cellule est ici basé sur la cellule définie dans le Modèle de Potts Cellulaire. Ce modèle a été étendu (MorphoPotts) pour permettre aux cellules de cibler une forme générique et dynamique et, de définir un bilan énergétique. La théorie du darwinisme cellulaire, une théorie originale de l’embryogenèse, a été simulée à l’aide du MorphoPotts. Des tissus ont émergé depuis une cellule œuf. Ces tissus sont cohérents car ils se renouvellent en continu et ont une forme reconnaissable. Pour vérifier si les interactions entre les MorphoPotts permettent au système de s’adapter et de s’auto-organiser, les performances des systèmes multi-agents ont dû être améliorées. Nous montrons que la programmation sur les cartes graphiques (GPU) amène à des gains de performance importants. Les simulations faites sur les cartes graphiques montrent comment un darwinisme cellulaire permet aux tissus cellulaires de s’auto-organiser et de s’adapter en réponse aux événements extérieurs
The self-organization between the cells gives an explanation of the cell tissues morphogenesis, like the phenomenon of embryogenesis. Meanwhile, there is no consensus on the nature of the interactions between the cells leading to this self-organization. On one hand, the in silico modelisation and simulation offers a formal support to help the biologist in his understanding of the phenomenon and gives arguments in favor of a theory or of another one. The computer science implementation of biological process allows, on the other hand, improving the existing computer science models. The multi-agent systems are computer science models which represent each entity (agent) of a system in an explicit way. The agents are executed in an autonomous way and in interaction with the others. Cell agents can be simulated with a multi-agent system, where interactions are based on the consumption and production of molecules, but also on the adhesion and pressure cells exert on each other. A cell agent is based on the cell defined in the Cellular Potts Model. This model has been extended (MorphoPotts) in order to allow the cells to reach a generic and dynamic shape and to define an energy balance. The theory of Darwin at cellular level, an original theory of the embryogenesis, has been simulated via MorphoPotts where tissues emerge from one stem cell. These tissues are coherent because they have a continue renewal and a recognizable shape. To verify if the interactions between the MorphoPotts allow the system to self-organize and self-adapt, the performances of the multi-agent systems were enhanced. We show that the graphics processing unit (GPU) programming leads to considerable performance gains. The simulations done on the OPU show that a cellular darwinism allows the cell tissue to self-organize and self-adapt in reply to exterior events

M.Sc. (Computer Science)
The purpose of the research presented in this dissertation is to explore the use of a distributed multi-agent system in a general purpose crowd simulation model. Crowd simulation is becoming an increasingly important tool for analysing new construction projects, as it enables safety and performance evaluations to be performed on architectural plans before the buildings have been constructed. Crowd simulation is a challenging problem, as it requires the simulation of complex interactions of people within a crowd. The dissertation investigates existing models of crowd simulation and identifies three primary sub-tasks of crowd simulation: deliberation, path planning and collision-avoiding movement. Deliberation is the process of determining which goal an agent will attempt to satisfy next. Path planning is the process of finding a collision-free path from an agent‟s current location towards its goal. Collision-avoiding movement deals with moving an agent along its calculated path while avoiding collisions with other agents. A multi-agent crowd simulation model, DiMACS, is proposed as a means of addressing the problem of crowd simulation. Multi-agent technology provides an effective solution for representing individuals within a crowd; each member of a crowd can be represented as an intelligent agent. Intelligent agents are capable of maintaining their own internal state and deciding on a course of action based on that internal state. DiMACS is capable of producing realistic simulations while making use of distributed and parallel processing to improve its performance. In addition, the model is highly customisable. The dissertation also presents a user-friendly method for configuring agents within a simulation that abstracts the complexity of agent behaviour away from a user so as to increase the accessibility of configuring the proposed model. In addition, an application programming interface is provided that enables developers to extend the model to simulate additional agent behaviours. The research shows how distributed and parallel processing may be used to improve the performance of an agent-based crowd simulation without compromising the accuracy of the simulation.

碩士
國立中興大學
化學系所
99
Chemical cross-linking of biological components is a recent outgrowth of chemical modification of proteins. The general approach for the labeling of proteins with small organic molecules involves the synthesis of a chemically reactive analog of the desired structure followed by reaction with the protein of interest. We design a reaction for conjugating the fluorescent molecule (2-naphthol) to a protein by several cross-linkers, such as 2-Iminothiolane (2-IT), p-Maleimidophenyl Isocynate (PMPI), succinic anhydride, and glutaric anhydride. We found that 2-IT is a useful coupling reagent, which is capable of reacting with the side chains of residues in a protein. The succinic anhydride and glutaric anhydride are used because they can be relatively easy to obtain, and can form the chemical bonding between the reactive amino acid side chains in proteins and fluorescent molecules.

Through careful examination of primary documents, this book reveals that the true purpose of Zebulon Pike's western expedition in 1806–1807 was not innocent exploration of the West but an espionage mission in preparation for an American invasion of New Mexico. In 1806, the United States was on the brink of war with Spain over the disputed western boundary of the Louisiana Purchase and anticipated invading New Mexico. Possessing only meager information on the terrain as well as Spanish troop numbers and fortifications, President Thomas Jefferson and General James Wilkinson needed an agent who could cross the continent to Spanish territory all the way to the capital of Santa Fe, provide a plausible "cover story" for being there, and return with the vital information. Zebulon Pike: Thomas Jefferson's Agent for Empire is the story of how Pike carried out his mission as a prelude to an American invasion of New Mexico. This unique book is the first to fully chronicle Pike's 1806–1807 expedition to the Rocky Mountains and beyond, addressing both the exploratory and the clandestine purposes of Pike's western journey. By carefully examining the evidence available in primary documents, which shows that Pike prepared and carried a map showing a route to Santa Fe on his journey, the author overturns the conventional wisdom that Pike became lost searching for the Red River and mistakenly wandered into Spanish territory. This book also presents the Spanish perspective and response to Pike's invasion of Spanish territory and provides historical context to understand the role of Pike's expedition in Thomas Jefferson's quest to build an American "empire for liberty."

This educational product is part of the author's dissertation carried out in the Postgraduate Program in Professional Master's in National Network for Teaching Environmental Sciences at the Federal University of Paraná. The work entitled Protocol for the application of Environmental Education activities for teachers of Rural Education originated from the activities developed at the Municipal School of Campo in Amparo, a fishing community, located in Paranaguá Bay, in the State of Paraná. The general objective of this investigation was to develop actions of awareness and Environmental Education in the School to be applied in the community. Based on these actions, the purpose was to develop a protocol for rural education teachers to train child environmental agents with the function of being multipliers of local knowledge and citizenship attitudes for their families and community residents. In the moments of interaction between school and community, we tried to work on awareness in environmental education so that in the future local environmental problems can be reduced or eliminated.

Most of Mexican populations, even medical population, do not have enough information on rheumatic diseases to allow them to understand the importance of the socioeconomic and psychological impact of these disorders. In order to reduce this problem, a didactic tool is provided to general practitioners that enables them to identify medical disorders in the area of rheumatology. Didactic software is based on the application of an intelligent agent based on goals that contains enough information to identify seven of the most common inflammatory rheumatic diseases and fourteen non-inflammatory. The purpose of this tool is that a general practitioner can get an early diagnosis in a rheumatic patient and subsequently send that patient to a rheumatologist in order to prevent damage. The presented prototype can be useful for professors and students of the computation area to solve similar problems.

Intelligent agents are becoming increasingly important in our society in applications as diverse as house cleaning robots, computer-controlled opponents in video games, unmanned aerial combat vehicles, entertainment robots, and autonomous explorers in outer space. However, the broader adoption of intelligent agents is often hindered by their limited adaptability to new tasks; when conditions change slightly, agents may quickly become confused. Additionally, a substantial engineering effort is required to design an agent for each new task. This paper presents an adaptable, general purpose intelligent agent toolkit based on reinforcement learning (RL), an approach with strong mathematical foundations and intriguing biological implications. RL algorithms are powerful because of their generality: agents simply receive a scalar reward value representing success or failure, which greatly simplifies the agent design process. Furthermore, these algorithms can be combined with other techniques (e.g., planning from a learned internal model) to improve learning efficiency. The design and implementation of an open source RL toolkit is presented here as a step towards the goal of general purpose agents. Experimental results show learning performance on several tasks, including two physical control problems.

The lack of systematic pain management training and support among primary care physicians (PCPs) limits their ability to provide quality care for patients with pain. Here, we demonstrate an Agent-based Clinical Decision Support System to empower PCPs to leverage knowledge from pain specialists. The system learns a general-purpose representation space on patients, automatically diagnoses pain, recommends therapy and medicine, and suggests a referral program to PCPs in their decision-making tasks.

In this work, we have obtained multi-purpose optimization of the baklava production process based on technologies of agent systems and the intelligent process control environment. The general architecture of the control system intelligent environment with agent technologies for recognizing abnormal situations is developed and adaptive baklava production regulators are synthesized. The proposed approach to the construction of an automatic process control system with an intelligent environment is to increase the efficiency of baklava production.

<p>This contribution focuses on the numerical modelling and simulation of the mechanical behaviour of structural elements containing a self-healing agent. Specifically, the finite element modelling of the mechanical response of plain concrete structural element, containing a healing agent and subject to various loading conditions is discussed. A customized numerical procedure designed to implement the Damage Mechanics-based constitutive relation for concrete into a general-purpose finite element code is developed. The procedure comprises algorithms for evaluation of the volume of newly-formed cracks, the recovery of stiffness of the structural element due to crack closure, the initiation of healing and its effect on the overall response of the structural element. The procedure is demonstrated by simulations of a concrete cylinder subjected to compression and torsion.</p>

<p>This contribution focuses on the numerical modelling and simulation of the mechanical behaviour of structural elements containing a self-healing agent. Specifically, the finite element modelling of the mechanical response of plain concrete structural element, containing a healing agent and subject to various loading conditions is discussed. A customized numerical procedure designed to implement the Damage Mechanics-based constitutive relation for concrete into a general-purpose finite element code is developed. The procedure comprises algorithms for evaluation of the volume of newly-formed cracks, the recovery of stiffness of the structural element due to crack closure, the initiation of healing and its effect on the overall response of the structural element. The procedure is demonstrated by simulations of a concrete cylinder subjected to compression and torsion.</p>

The purpose of this document is to describe a version of the Global Trade Analysis Project (GTAP) computable general equilibrium (CGE) model implemented in the General Algebraic Modeling System (GAMS) that is a literal implementation of the standard General Economic Modeling Package (GEMPACK) version. It updates and supersedes the model description provided in van der Mensbrugghe (2018), which remains the main reference. The key revision of the new version is that the model is calibrated to initially normalized variables, similar to Rutherford’s GTAPinGams model (Lanz and Rutherford (2016)). One objective is to improve model scaling and hopefully numerical convergence. One advantage is that it is relatively easy to view the results in percentage change form, similar to the GEMPACK version of the model. The new version includes three additional extensions: national-level Armington preferences for domestic vs. imports, agent-level sourcing by region of origin (a variant of the multi-region input-output model (MRIO)), and rent-generating ad valorem equivalent (AVE) of non-tariff measures (NTMs).

Purpose/Background: Postoperative nausea and vomiting (PONV) is a frequent complaint in the postoperative period, which can delay discharge, result in readmission, and increase cost for patients and facilities. Inducing paralysis is common in anesthesia, as is utilizing the drugs neostigmine and sugammadex as reversal agents for non-depolarizing neuromuscular blockers. Many studies are available that compare these two drugs to determine if neostigmine increases the risk of PONV over sugammadex. Sugammadex has a more favorable pharmacologic profile and may improve patient outcomes by reducing PONV. Methods: This review included screening a total of 39 studies and peer-reviewed articles that looked at patients undergoing general anesthesia who received non-depolarizing neuromuscular blockers requiring either neostigmine or sugammadex for reversal, along with their respective PONV rates. 8 articles were included, while 31 articles were removed based on our exclusion criteria. These were published between 2014 and 2020 exclusively. The key words used were “neostigmine”, “sugammadex”, “PONV”, along with combinations “paralytic reversal agents and PONV”. This search was performed on the scholarly database MEDLINE. The data items were PONV rates in neostigmine group, PONV rates in sugammadex group, incidence of postoperative analgesic consumption in neostigmine group, and incidence of postoperative analgesic consumption in sugammadex group. Results: Despite numerical differences being noted in the incidence of PONV with sugammadex over reversal with neostigmine, there did not appear to be any statistically significant data in the multiple peer-reviewed trials included in our review, for not one of the 8 studies concluded that there was a higher incidence of PONV in one drug or the other of an y clinical relevance. Although the side-effect profile tended to be better in the sugammadex group than neostigmine in areas other than PONV, there was not sufficient evidence to conclude that one drug was superior to the other in causing a direct reduction of PONV. Implications for Nursing Practice: There were variable but slight differences noted between both drug groups in PONV rates, but it remained that none of the studies determined it was statically significant or clinically conclusive. This review did, however, note other advantages to sugammadex over neostigmine, including its pharmacologic profile of more efficiently reversing non-depolarizing neuromuscular blocking drugs and its more favorable pharmacokinetics. This lack of statistically significant evidence found within these studies consequentially does not support pharmacologic decision-making of one drug in favor of the other for reducing PONV; therefore, PONV alone is not a sufficient rationale for a provider to justify using one reversal over another at the current time until further research proves otherwise.

Objectives: The overall objective as defined in the approved proposal was the development of a whole-cell sensor panel for the detection of endocrine disruption activities of agriculturally relevant chemicals. To achieve this goal several specific objectives were outlined: (a) The development of new genetically engineered wholecell sensor strains; (b) the combination of multiple strains into a single sensor panel to effect multiple response modes; (c) development of a computerized algorithm to analyze the panel responses; (d) laboratory testing and calibration; (e) field testing. In the course of the project, mostly due to the change in the US partner, three modifications were introduced to the original objectives: (a) the scope of the project was expanded to include pharmaceuticals (with a focus on antibiotics) in addition to endocrine disrupting chemicals, (b) the computerized algorithm was not fully developed and (c) the field test was not carried out. Background: Chemical agents, such as pesticides applied at inappropriate levels, may compromise water quality or contaminate soils and hence threaten human populations. In recent years, two classes of compounds have been increasingly implicated as emerging risks in agriculturally-related pollution: endocrine disrupting compounds (EDCs) and pharmaceuticals. The latter group may reach the environment by the use of wastewater effluents, whereas many pesticides have been implicated as EDCs. Both groups pose a threat in proportion to their bioavailability, since that which is biounavailable or can be rendered so is a priori not a threat; bioavailability, in turn, is mediated by complex matrices such as soils. Genetically engineered biosensor bacteria hold great promise for sensing bioavailability because the sensor is a live soil- and water-compatible organism with biological response dynamics, and because its response can be genetically “tailored” to report on general toxicity, on bioavailability, and on the presence of specific classes of toxicants. In the present project we have developed a bacterial-based sensor panel incorporating multiple strains of genetically engineered biosensors for the purpose of detecting different types of biological effects. The overall objective as defined in the approved proposal was the development of a whole-cell sensor panel for the detection of endocrine disruption activities of agriculturally relevant chemicals. To achieve this goal several specific objectives were outlined: (a) The development of new genetically engineered wholecell sensor strains; (b) the combination of multiple strains into a single sensor panel to effect multiple response modes; (c) development of a computerized algorithm to analyze the panel responses; (d) laboratory testing and calibration; (e) field testing. In the course of the project, mostly due to the change in the US partner, three modifications were introduced to the original objectives: (a) the scope of the project was expanded to include pharmaceuticals (with a focus on antibiotics) in addition to endocrine disrupting chemicals, (b) the computerized algorithm was not fully developed and (c) the field test was not carried out. Major achievements: (a) construction of innovative bacterial sensor strains for accurate and sensitive detection of agriculturally-relevant pollutants, with a focus on endocrine disrupting compounds (UK and HUJ) and antibiotics (HUJ); (b) optimization of methods for long-term preservation of the reporter bacteria, either by direct deposition on solid surfaces (HUJ) or by the construction of spore-forming Bacillus-based sensors (UK); (c) partial development of a computerized algorithm for the analysis of sensor panel responses. Implications: The sensor panel developed in the course of the project was shown to be applicable for the detection of a broad range of antibiotics and EDCs. Following a suitable development phase, the panel will be ready for testing in an agricultural environment, as an innovative tool for assessing the environmental impacts of EDCs and pharmaceuticals. Furthermore, while the current study relates directly to issues of water quality and soil health, its implications are much broader, with potential uses is risk-based assessment related to the clinical, pharmaceutical, and chemical industries as well as to homeland security.

Banco de la República is celebrating its 100th anniversary in 2023. This is a very significant anniversary and one that provides an opportunity to highlight the contribution the Bank has made to the country’s development. Its track record as guarantor of monetary stability has established it as the one independent state institution that generates the greatest confidence among Colombians due to its transparency, management capabilities, and effective compliance with the central banking and cultural responsibilities entrusted to it by the Constitution and the Law. On a date as important as this, the Board of Directors of Banco de la República (BDBR) pays tribute to the generations of governors and officers whose commitment and dedication have contributed to the growth of this institution.1 Banco de la República’s mandate was confirmed in the National Constitutional Assembly of 1991 where the citizens had the opportunity to elect the seventy people who would have the task of drafting a new constitution. The leaders of the three political movements with the most votes were elected as chairs to the Assembly, and this tripartite presidency reflected the plurality and the need for consensus among the different political groups to move the reform forward. Among the issues considered, the National Constitutional Assembly gave special importance to monetary stability. That is why they decided to include central banking and to provide Banco de la República with the necessary autonomy to use the instruments for which they are responsible without interference from other authorities. The constituent members understood that ensuring price stability is a state duty and that the entity responsible for this task must be enshrined in the Constitution and have the technical capability and institutional autonomy necessary to adopt the decisions they deem appropriate to achieve this fundamental objective in coordination with the general economic policy. In particular, Article 373 established that “the State, through Banco de la República, shall ensure the maintenance of the purchasing power of the currency,” a provision that coincided with the central banking system adopted by countries that have been successful in controlling inflation. In 1999, in Ruling 481, the Constitutional Court stated that “the duty to maintain the purchasing power of the currency applies to not only the monetary, credit, and exchange authority, i.e., the Board of Banco de la República, but also those who have responsibilities in the formulation and implementation of the general economic policy of the country” and that “the basic constitutional purpose of Banco de la República is the protection of a sound currency. However, this authority must take the other economic objectives of state intervention such as full employment into consideration in their decisions since these functions must be coordinated with the general economic policy.” The reforms to Banco de la República agreed upon in the Constitutional Assembly of 1991 and in Act 31/1992 can be summarized in the following aspects: i) the Bank was assigned a specific mandate: to maintain the purchasing power of the currency in coordination with the general economic policy; ii) the BDBR was designatedas the monetary, foreign exchange, and credit authority; iii) the Bank and its Board of Directors were granted a significant degree of independence from the government; iv) the Bank was prohibited from granting credit to the private sector except in the case of the financial sector; v) established that in order to grant credit to the government, the unanimous vote of its Board of Directors was required except in the case of open market transactions; vi) determined that the legislature may, in no case, order credit quotas in favor of the State or individuals; vii) Congress was appointed, on behalf of society, as the main addressee of the Bank’s reporting exercise; and viii) the responsibility for inspection, surveillance, and control over Banco de la República was delegated to the President of the Republic. The members of the National Constitutional Assembly clearly understood that the benefits of low and stable inflation extend to the whole of society and contribute mto the smooth functioning of the economic system. Among the most important of these is that low inflation promotes the efficient use of productive resources by allowing relative prices to better guide the allocation of resources since this promotes economic growth and increases the welfare of the population. Likewise, low inflation reduces uncertainty about the expected return on investment and future asset prices. This increases the confidence of economic agents, facilitates long-term financing, and stimulates investment. Since the low-income population is unable to protect itself from inflation by diversifying its assets, and a high proportion of its income is concentrated in the purchase of food and other basic goods that are generally the most affected by inflationary shocks, low inflation avoids arbitrary redistribution of income and wealth.2 Moreover, low inflation facilitates wage negotiations, creates a good labor climate, and reduces the volatility of employment levels. Finally, low inflation helps to make the tax system more transparent and equitable by avoiding the distortions that inflation introduces into the value of assets and income that make up the tax base. From the monetary authority’s point of view, one of the most relevant benefits of low inflation is the credibility that economic agents acquire in inflation targeting, which turns it into an effective nominal anchor on price levels. Upon receiving its mandate, and using its autonomy, Banco de la República began to announce specific annual inflation targets as of 1992. Although the proposed inflation targets were not met precisely during this first stage, a downward trend in inflation was achieved that took it from 32.4% in 1990 to 16.7% in 1998. At that time, the exchange rate was kept within a band. This limited the effectiveness of monetary policy, which simultaneously sought to meet an inflation target and an exchange rate target. The Asian crisis spread to emerging economies and significantly affected the Colombian economy. The exchange rate came under strong pressure to depreciate as access to foreign financing was cut off under conditions of a high foreign imbalance. This, together with the lack of exchange rate flexibility, prevented a countercyclical monetary policy and led to a 4.2% contraction in GDP that year. In this context of economic slowdown, annual inflation fell to 9.2% at the end of 1999, thus falling below the 15% target set for that year. This episode fully revealed how costly it could be, in terms of economic activity, to have inflation and exchange rate targets simultaneously. Towards the end of 1999, Banco de la República announced the adoption of a new monetary policy regime called the Inflation Targeting Plan. This regime, known internationally as ‘Inflation Targeting,’ has been gaining increasing acceptance in developed countries, having been adopted in 1991 by New Zealand, Canada, and England, among others, and has achieved significant advances in the management of inflation without incurring costs in terms of economic activity. In Latin America, Brazil and Chile also adopted it in 1999. In the case of Colombia, the last remaining requirement to be fulfilled in order to adopt said policy was exchange rate flexibility. This was realized around September 1999, when the BDBR decided to abandon the exchange-rate bands to allow the exchange rate to be freely determined in the market.Consistent with the constitutional mandate, the fundamental objective of this new policy approach was “the achievement of an inflation target that contributes to maintaining output growth around its potential.”3 This potential capacity was understood as the GDP growth that the economy can obtain if it fully utilizes its productive resources. To meet this objective, monetary policy must of necessity play a countercyclical role in the economy. This is because when economic activity is below its potential and there are idle resources, the monetary authority can reduce the interest rate in the absence of inflationary pressure to stimulate the economy and, when output exceeds its potential capacity, raise it. This policy principle, which is immersed in the models for guiding the monetary policy stance, makes the following two objectives fully compatible in the medium term: meeting the inflation target and achieving a level of economic activity that is consistent with its productive capacity. To achieve this purpose, the inflation targeting system uses the money market interest rate (at which the central bank supplies primary liquidity to commercial banks) as the primary policy instrument. This replaced the quantity of money as an intermediate monetary policy target that Banco de la República, like several other central banks, had used for a long time. In the case of Colombia, the objective of the new monetary policy approach implied, in practical terms, that the recovery of the economy after the 1999 contraction should be achieved while complying with the decreasing inflation targets established by the BDBR. The accomplishment of this purpose was remarkable. In the first half of the first decade of the 2000s, economic activity recovered significantly and reached a growth rate of 6.8% in 2006. Meanwhile, inflation gradually declined in line with inflation targets. That was how the inflation rate went from 9.2% in 1999 to 4.5% in 2006, thus meeting the inflation target established for that year while GDP reached its potential level. After this balance was achieved in 2006, inflation rebounded to 5.7% in 2007, above the 4.0% target for that year due to the fact that the 7.5% GDP growth exceeded the potential capacity of the economy.4 After proving the effectiveness of the inflation targeting system in its first years of operation, this policy regime continued to consolidate as the BDBR and the technical staff gained experience in its management and state-of-the-art economic models were incorporated to diagnose the present and future state of the economy and to assess the persistence of inflation deviations and expectations with respect to the inflation target. Beginning in 2010, the BDBR established the long-term 3.0% annual inflation target, which remains in effect today. Lower inflation has contributed to making the macroeconomic environment more stable, and this has favored sustained economic growth, financial stability, capital market development, and the functioning of payment systems. As a result, reductions in the inflationary risk premia and lower TES and credit interest rates were achieved. At the same time, the duration of public domestic debt increased significantly going from 2.27 years in December 2002 to 5.86 years in December 2022, and financial deepening, measured as the level of the portfolio as a percentage of GDP, went from around 20% in the mid-1990s to values above 45% in recent years in a healthy context for credit institutions.Having been granted autonomy by the Constitution to fulfill the mandate of preserving the purchasing power of the currency, the tangible achievements made by Banco de la República in managing inflation together with the significant benefits derived from the process of bringing inflation to its long-term target, make the BDBR’s current challenge to return inflation to the 3.0% target even more demanding and pressing. As is well known, starting in 2021, and especially in 2022, inflation in Colombia once again became a serious economic problem with high welfare costs. The inflationary phenomenon has not been exclusive to Colombia and many other developed and emerging countries have seen their inflation rates move away from the targets proposed by their central banks.5 The reasons for this phenomenon have been analyzed in recent Reports to Congress, and this new edition delves deeper into the subject with updated information. The solid institutional and technical base that supports the inflation targeting approach under which the monetary policy strategy operates gives the BDBR the necessary elements to face this difficult challenge with confidence. In this regard, the BDBR reiterated its commitment to the 3.0% inflation target in its November 25 communiqué and expects it to be reached by the end of 2024.6 Monetary policy will continue to focus on meeting this objective while ensuring the sustainability of economic activity, as mandated by the Constitution. Analyst surveys done in March showed a significant increase (from 32.3% in January to 48.5% in March) in the percentage of responses placing inflation expectations two years or more ahead in a range between 3.0% and 4.0%. This is a clear indication of the recovery of credibility in the medium-term inflation target and is consistent with the BDBR’s announcement made in November 2022. The moderation of the upward trend in inflation seen in January, and especially in February, will help to reinforce this revision of inflation expectations and will help to meet the proposed targets. After reaching 5.6% at the end of 2021, inflation maintained an upward trend throughout 2022 due to inflationary pressures from both external sources, associated with the aftermath of the pandemic and the consequences of the war in Ukraine, and domestic sources, resulting from: strengthening of local demand; price indexation processes stimulated by the increase in inflation expectations; the impact on food production caused by the mid-2021 strike; and the pass-through of depreciation to prices. The 10% increase in the minimum wage in 2021 and the 16% increase in 2022, both of which exceeded the actual inflation and the increase in productivity, accentuated the indexation processes by establishing a high nominal adjustment benchmark. Thus, total inflation went to 13.1% by the end of 2022. The annual change in food prices, which went from 17.2% to 27.8% between those two years, was the most influential factor in the surge in the Consumer Price Index (CPI). Another segment that contributed significantly to price increases was regulated products, which saw the annual change go from 7.1% in December 2021 to 11.8% by the end of 2022. The measure of core inflation excluding food and regulated items, in turn, went from 2.5% to 9.5% between the end of 2021 and the end of 2022. The substantial increase in core inflation shows that inflationary pressure has spread to most of the items in the household basket, which is characteristic of inflationary processes with generalized price indexation as is the case in Colombia. Monetary policy began to react early to this inflationary pressure. Thus, starting with its September 2021 session, the BDBR began a progressive change in the monetary policy stance moving away from the historical low of a 1.75% policy rate that had intended to stimulate the recovery of the economy. This adjustment process continued without interruption throughout 2022 and into the beginning of 2023 when the monetary policy rate reached 12.75% last January, thus accumulating an increase of 11 percentage points (pp). The public and the markets have been surprised that inflation continued to rise despite significant interest rate increases. However, as the BDBR has explained in its various communiqués, monetary policy works with a lag. Just as in 2022 economic activity recovered to a level above the pre-pandemic level, driven, along with other factors, by the monetary stimulus granted during the pandemic period and subsequent months, so too the effects of the current restrictive monetary policy will gradually take effect. This will allow us to expect the inflation rate to converge to 3.0% by the end of 2024 as is the BDBR’s purpose.Inflation results for January and February of this year showed declining marginal increases (13 bp and 3 bp respectively) compared to the change seen in December (59 bp). This suggests that a turning point in the inflation trend is approaching. In other Latin American countries such as Chile, Brazil, Perú, and Mexico, inflation has peaked and has begun to decline slowly, albeit with some ups and downs. It is to be expected that a similar process will take place in Colombia in the coming months. The expected decline in inflation in 2023 will be due, along with other factors, to lower cost pressure from abroad as a result of the gradual normalization of supply chains, the overcoming of supply shocks caused by the weather, and road blockades in previous years. This will be reflected in lower adjustments in food prices, as has already been seen in the first two months of the year and, of course, the lagged effect of monetary policy. The process of inflation convergence to the target will be gradual and will extend beyond 2023. This process will be facilitated if devaluation pressure is reversed. To this end, it is essential to continue consolidating fiscal sustainability and avoid messages on different public policy fronts that generate uncertainty and distrust. 1 This Report to Congress includes Box 1, which summarizes the trajectory of Banco de la República over the past 100 years. In addition, under the Bank’s auspices, several books that delve into various aspects of the history of this institution have been published in recent years. See, for example: Historia del Banco de la República 1923-2015; Tres banqueros centrales; Junta Directiva del Banco de la República: grandes episodios en 30 años de historia; Banco de la República: 90 años de la banca central en Colombia. 2 This is why lower inflation has been reflected in a reduction of income inequality as measured by the Gini coefficient that went from 58.7 in 1998 to 51.3 in the year prior to the pandemic. 3 See Gómez Javier, Uribe José Darío, Vargas Hernando (2002). “The Implementation of Inflation Targeting in Colombia”. Borradores de Economía, No. 202, March, available at: https://repositorio.banrep.gov.co/handle/20.500.12134/5220 4 See López-Enciso Enrique A.; Vargas-Herrera Hernando and Rodríguez-Niño Norberto (2016). “The inflation targeting strategy in Colombia. An historical view.” Borradores de Economía, No. 952. https://repositorio.banrep.gov.co/handle/20.500.12134/6263 5 According to the IMF, the percentage change in consumer prices between 2021 and 2022 went from 3.1% to 7.3% for advanced economies, and from 5.9% to 9.9% for emerging market and developing economies. 6 https://www.banrep.gov.co/es/noticias/junta-directiva-banco-republica-reitera-meta-inflacion-3