Marco Lagi
Cambridge, Massachusetts, United States
750 followers
500+ connections
About
Machine learning leader, engineer, and researcher. Published more than 40 papers in 5…
Experience
Education
Licenses & Certifications
Publications
-
Accurate market price formation model with both supply-demand and trend-following for global food prices providing policy recommendations
Proceedings of the National Academy of Sciences
Recent increases in food prices are linked to widespread hunger and social unrest. The causes of high food prices have been debated. Here we rule out explanations that are not consistent with the data and construct a dynamic model of food prices using two factors determined to have the largest impact: corn-to-ethanol conversion and investor speculation. We overcome limitations of equilibrium theories that are unable to quantify the impact of speculation by using a dynamic model of trend…
Recent increases in food prices are linked to widespread hunger and social unrest. The causes of high food prices have been debated. Here we rule out explanations that are not consistent with the data and construct a dynamic model of food prices using two factors determined to have the largest impact: corn-to-ethanol conversion and investor speculation. We overcome limitations of equilibrium theories that are unable to quantify the impact of speculation by using a dynamic model of trend following. The model accurately fits the data. Ethanol conversion results in a smooth price increase, whereas speculation results in bubbles and crashes. These findings significantly inform the discussion about food prices and market equilibrium and have immediate policy implications.
-
Anticipating Economic Market Crises Using Measures of Collective Panic
PloS one
Predicting panic is of critical importance in many areas of human and animal behavior, notably in the context of economics. The recent financial crisis is a case in point. Panic may be due to a specific external threat or self-generated nervousness. Here we show that the recent economic crisis and earlier large single-day panics were preceded by extended periods of high levels of market mimicry—direct evidence of uncertainty and nervousness, and of the comparatively weak influence of external…
Predicting panic is of critical importance in many areas of human and animal behavior, notably in the context of economics. The recent financial crisis is a case in point. Panic may be due to a specific external threat or self-generated nervousness. Here we show that the recent economic crisis and earlier large single-day panics were preceded by extended periods of high levels of market mimicry—direct evidence of uncertainty and nervousness, and of the comparatively weak influence of external news. High levels of mimicry can be a quite general indicator of the potential for self-organized crises.
-
Food Security and Political Instability: From Ethanol and Speculation to Riots and Revolutions
Conflict and Complexity/Springer
Social unrest may reflect a variety of factors such as poverty, unemployment, and social injustice. Despite the many possible contributing factors, the timing of violent protests in North Africa and the Middle East in 2011 as well as earlier riots in 2008 coincides with large peaks in global food prices. We identify a specific food price threshold above which protests become likely. These observations suggest that protests may reflect not only long-standing political failings of governments but…
Social unrest may reflect a variety of factors such as poverty, unemployment, and social injustice. Despite the many possible contributing factors, the timing of violent protests in North Africa and the Middle East in 2011 as well as earlier riots in 2008 coincides with large peaks in global food prices. We identify a specific food price threshold above which protests become likely. These observations suggest that protests may reflect not only long-standing political failings of governments but also the sudden desperate straits of vulnerable populations. If food prices remain high, there is likely to be persistent and increasing global social disruption. Underlying the food price peaks we also find an ongoing trend of increasing prices. We extrapolate these trends and identify a crossing point to the domain of high impacts, even without price peaks, in 2012–2013. This implies that avoiding global food crises and associated social unrest requires rapid and concerted action.
-
South African riots: repercussion of the global food crisis and US drought
Conflict and Complexity/Springer
High and volatile global food prices have led to food riots and played a critical role in triggering the Arab Spring revolutions in recent years. The severe drought in the USA in the summer of 2012 led to a new increase in food prices. Through the fall, they remained at a threshold above which the riots and revolutions had predominantly occurred. Global prices at this level create conditions where an exacerbating local circumstance can trigger unrest. Global corn (maize) prices reached new…
High and volatile global food prices have led to food riots and played a critical role in triggering the Arab Spring revolutions in recent years. The severe drought in the USA in the summer of 2012 led to a new increase in food prices. Through the fall, they remained at a threshold above which the riots and revolutions had predominantly occurred. Global prices at this level create conditions where an exacerbating local circumstance can trigger unrest. Global corn (maize) prices reached new highs, and countries that depend mostly on maize are more likely to experience high local food prices and associated pressures toward social unrest. Here we analyze the conditions in South Africa, which is a heavily maize-dependent country, and we find that consumer food indices have increased dramatically. Coinciding with the food price increases this summer, massive labor strikes in mining and agriculture have led to the greatest single incident of social violence since the fall of apartheid in 1994. Worker demands for dramatic pay increases reflect that their wages have not kept up with drastic increases in the prices of necessities, especially food. Without attention to the global food price situation, more incidents of food-based social instability are likely to arise. Other countries that have manifested food-related protests and riots in 2012 include Haiti (prior to the impact of Hurricane Sandy) and Argentina. Moreover, these cases of unrest are just the most visible symptom of widespread suffering of poor populations worldwide due to elevated food prices. Our analysis has shown that policy decisions that would directly impact food prices are decreasing the conversion of maize to ethanol in the USA, and reimposing regulations on commodity futures markets to prevent excessive speculation, which we have shown causes bubbles and crashes in these markets.
-
Economics of food prices and crises
NECSI
Recent increases in basic food prices are severely impacting vulnerable populations worldwide. Proposed causes such as shortages of grain due to adverse weather, increasing meat consumption in China and India, conversion of corn to ethanol in the US, and investor speculation on commodity markets lead to widely differing implications for policy. A lack of clarity about which factors are responsible reinforces policy inaction. Here, for the first time, we construct a dynamic model that…
Recent increases in basic food prices are severely impacting vulnerable populations worldwide. Proposed causes such as shortages of grain due to adverse weather, increasing meat consumption in China and India, conversion of corn to ethanol in the US, and investor speculation on commodity markets lead to widely differing implications for policy. A lack of clarity about which factors are responsible reinforces policy inaction. Here, for the first time, we construct a dynamic model that quantitatively agrees with food prices. The results show that the dominant causes of price increases are investor speculation and ethanol conversion. Models that just treat supply and demand are not consistent with the actual price dynamics. The two sharp peaks in 2007/2008 and 2010/2011 are specifically due to investor speculation, while an underlying upward trend is due to increasing demand from ethanol conversion. The model includes investor trend following as well as shifting between commodities, equities and bonds to take advantage of increased expected returns. Claims that speculators cannot influence grain prices are shown to be invalid by direct analysis of price setting practices of granaries. Both causes of price increase, speculative investment and ethanol conversion, are promoted by recent regulatory changes|deregulation of the commodity markets, and policies promoting the conversion of corn to ethanol. Rapid action is needed to reduce the impacts
of the price increases on global hunger. -
Economics of Collective Value
NECSI
Price and value are not synonymous, they are related concepts. Value is determined by desire of an individual for a good, price is determined by collective agreements on transactions to transfer that good. Such a shift between scales is at the heart of complex systems research, which quantifies the emergence of larger scale phenomena from smaller scale interactions. Attempts have been made toward formalizing theories of price formation, but none of them so far was able to reproduce historical…
Price and value are not synonymous, they are related concepts. Value is determined by desire of an individual for a good, price is determined by collective agreements on transactions to transfer that good. Such a shift between scales is at the heart of complex systems research, which quantifies the emergence of larger scale phenomena from smaller scale interactions. Attempts have been made toward formalizing theories of price formation, but none of them so far was able to reproduce historical economic data in a quantitative way.
-
Preuve de la manipulation des marches lors de la crise financiere
NECSI
Nous présentons la preuve directe de la manipulation des marchés au début de la crise financière en novembre 2007. Le type spécifique de manipulation, dénommé «le raid baissier» aurait été empêché par une règle que l'autorité du Marché, la Commission sur la Sécurité et les échanges [le « SEC »] a annulé au mois de juillet 2007. Cette règle, dite « clôture à la hausse » [« the uptick rule »] avait été formulée afin d’empêcher la manipulation des marchés et promouvoir la stabilité et a été…
Nous présentons la preuve directe de la manipulation des marchés au début de la crise financière en novembre 2007. Le type spécifique de manipulation, dénommé «le raid baissier» aurait été empêché par une règle que l'autorité du Marché, la Commission sur la Sécurité et les échanges [le « SEC »] a annulé au mois de juillet 2007. Cette règle, dite « clôture à la hausse » [« the uptick rule »] avait été formulée afin d’empêcher la manipulation des marchés et promouvoir la stabilité et a été appliquée comme une partie intégrante de la riposte du gouvernement au krach de 1928 et ses suites depuis 1938. Le 1er novembre 2007, il y a eu une hausse exceptionnelle du volume accompagnée de la chute du prix des titres de Citigroup. Notre analyse des données dont dispose l'industrie financière démontre que cette baisse correspondait à une augmentation hors norme de titres prêtées dont la vente aura fourni la plus grande portion du volume des transactions. La vente de titres empruntées ne peut s'expliquer dans le cadre de l’événementiel puisqu'il n'y a pas eu d'augmentation de ventes chez les porteurs. Une quantité similaire de titres ont été rendus lors d'une seule journée six jours plus tard. L'ordre de grandeur et la coordination entre l'emprunt et le retour des titres fournit la preuve d'un effort concerté de forcer la baisse du prix du titre de Citigroup et d'en tirer un profit, i.e. : un raid baissier Les interprétations et analyses des marchés financiers doivent prendre en considération la possibilité que les interventions d'acteurs individuels ou de groupes coordonnés peuvent faire fléchir les marchés. Les marchés ne sont suffisamment transparents ni pour exposer ni pour prévenir même des manipulations grossières. Nos résultats indiquent un besoin de réglementations qui empêchent des interventions concertés qui opèrent une fléchissement de la valeur d'équilibre et qui contribuent aux effondrements du marché.
-
Protein Dynamics Studied by Quasi-elastic Neutron Scattering
American Physical Society
The biological function and activities of proteins are intimately related to their structures and dynamics. Nowadays, neutron scattering is one of the most powerful tools to study the protein dynamics. In this study, we use quasielastic neutron scattering (QENS) at the Spallation Neutron Source, ORNL, to study relaxational dynamics of two structurally different proteins --- hen egg white lysozyme and an inorganic pyrophosphatase from a hyperthermophile, in the time range of 10ps to 1ns. We…
The biological function and activities of proteins are intimately related to their structures and dynamics. Nowadays, neutron scattering is one of the most powerful tools to study the protein dynamics. In this study, we use quasielastic neutron scattering (QENS) at the Spallation Neutron Source, ORNL, to study relaxational dynamics of two structurally different proteins --- hen egg white lysozyme and an inorganic pyrophosphatase from a hyperthermophile, in the time range of 10ps to 1ns. We experimentally prove that the slow dynamics of globular proteins can be described by the mode-coupling theory (MCT) that was originally developed for glass-forming molecular liquids. The MCT predicts the appearance of a logarithmic decay for a glass-forming liquid. Such dynamic behavior is also observed by recent molecular dynamics (MD) simulations on protein molecules. In addition, we compare the temperature dependence of the dynamics of the two proteins with completely different activity profiles. Our results greatly help understanding the relation between protein dynamics and their biological functions.
-
Evidence of market manipulation in the financial crisis
arXiv:1112.3095v1
-
Dynamics of a globular protein and its hydration water studied by neutron scattering and MD simulations
Spectroscopy 24, 1
-
Experimental Evidence of Logarithmic Relaxation in Single-particle Dynamics of Hydrated Protein Molecules
Soft Matter 6, 262
-
The Dynamic Response Function Chi_T fo Confined Supercooled Water and its Relation to the Dynamic Crossover Phenomenon
Z. Phys. Chem. 224, 109
-
Absence of the Density Minimum of Supercooled Water in Hydrophobic Confinement
J. Phys. Chem. B 113, 5007
-
Dynamic Susceptibility of Supercooled Water and its relation to the Dynamic Crossover Phenomenon
Phys. Rev. E 79, 040201
-
Dynamical Coupling between a Globular Protein and its Hydration Water Studied by Neutron Scattering and MD Simulation
WPI-AIMR-2009 Proceedings
-
Evidence of Dynamic Crossover Phenomena in Water and Other Glass-Forming Liquids: Experiments, MD Simulations and Theory
J. Phys.: Cond. Matt. 21, 504102
-
Logarithmic decay in single-particle relaxation of hydrated lysozyme powder
Phys. Rev. Lett. 103, 108102
-
Neutron Scattering Studies of Dynamic Crossover Phenomena in a Coupled System of Biopolymer and Its Hydration Water
Journal of Physics: Conference Series [JPCS] 177, 012006
-
Observation of high-temperature dynamic crossover in protein hydration water and its relation to reversible denaturation of lysozyme
J. Chem. Phys. 130, 135101
-
Collective headgroup conformational transition in twisted micellar superstructures
Soft Matter 4, 1102-1113
Predictions on amphiphilic self-assemblies traditionally rely on considerations on molecular shape and charge of the surfactant. In the case of functional surfactants a more sophisticated toolbox becomes necessary to design amphiphiles encoding chemical functionalities that provide additional responsive properties to their self-assemblies. Here we report on a comprehensive and combined structural–spectroscopic characterization of 1,2-dilauroyl-phosphatidyl-adenosine (DLPA) micelles in phosphate…
Predictions on amphiphilic self-assemblies traditionally rely on considerations on molecular shape and charge of the surfactant. In the case of functional surfactants a more sophisticated toolbox becomes necessary to design amphiphiles encoding chemical functionalities that provide additional responsive properties to their self-assemblies. Here we report on a comprehensive and combined structural–spectroscopic characterization of 1,2-dilauroyl-phosphatidyl-adenosine (DLPA) micelles in phosphate buffer. The temperature dependence, more precisely the thermal history of the sample, is explicitly taken into account. The experimental data, supplemented with MD simulations, indicate the presence of two possible states at room temperature, characterized by distinctly different structural properties that depend on the thermal history of the sample. The twisted superstructures, produced by aging DLPA micelles through intermicellar assembly of locally cylindrical aggregates at room temperature, collapse upon warming at 35 °C, yielding aligned filaments and/or wormlike structures. The initial superstructures cannot be recovered by thermal inversion. The reason for this behaviour is that the thermal activation causes a redistribution of syn–anti conformations of adenosine headgroups, as indicated by spectroscopic results (NMR, CD, FTIR), which is then collectively frozen thanks to molecular constraints present in the aggregate.
-
DNA Closed Nanostructures: a Structural and Monte Carlo Simulation Study
J. Phys. Chem. B 112, 15283
DNA nanoconstructs are obtained in solution by using six unique 42-mer DNA oligonucleotides, whose sequences have been designed to form a pseudohexagonal structure. The required flexibility is provided by the insertion of two non-base-paired thymines in the middle of each sequence that work as flexible hinges and constitute the corners of the nanostructure when formed. We show that hexagonally shaped nanostructures of about 7 nm diameter and their corresponding linear open constructs are formed…
DNA nanoconstructs are obtained in solution by using six unique 42-mer DNA oligonucleotides, whose sequences have been designed to form a pseudohexagonal structure. The required flexibility is provided by the insertion of two non-base-paired thymines in the middle of each sequence that work as flexible hinges and constitute the corners of the nanostructure when formed. We show that hexagonally shaped nanostructures of about 7 nm diameter and their corresponding linear open constructs are formed by self-assembly of the specifically designed linear oligonucleotides. The structural and dynamical characterization of the nanostructure is obtained in situ for the first time by using dynamic light scattering (DLS), a noninvasive method that provides a fast dynamic and structural analysis and allows the characterization of the different synthetic DNA nanoconstructs in solution. A validation of the LS results is obtained through Monte Carlo (MC) simulations and atomic force microscopy (AFM). In particular, a mesoscale molecular model for DNA, developed by Knotts et al., is exploited to perform MC simulations and to obtain information about the conformations as well as the conformational flexibilities of these nanostructures, while AFM provides a very detailed particle analysis that yields an estimation of the particle size and size distribution. The structural features obtained by MC and AFM are in good agreement with DLS, showing that DLS is a fast and reliable tool for characterization of DNA nanostructures in solution.
-
Interconnected Networks: Structural and Dynamic Characterization of Aqueous Dispersions of Dioctanoylphosphatidylcholine
J. Phys. Chem. B 112, 12625-12634
Aqueous dispersions of the phospholipid dioctanoylphosphatidylcholine (diC8PC) phase-separate below a cloud-point temperature, depending on lipid concentration. The lower phase is viscous and rich in lipid. The structure and dynamics of this system were explored via cryo-transmission electron microscopy (cryo-TEM), small-angle X-ray scattering (SAXS), and NMR. The lower phase comprises a highly interconnected tridimensional network of wormlike micelles. A molecular mechanism for the phase…
Aqueous dispersions of the phospholipid dioctanoylphosphatidylcholine (diC8PC) phase-separate below a cloud-point temperature, depending on lipid concentration. The lower phase is viscous and rich in lipid. The structure and dynamics of this system were explored via cryo-transmission electron microscopy (cryo-TEM), small-angle X-ray scattering (SAXS), and NMR. The lower phase comprises a highly interconnected tridimensional network of wormlike micelles. A molecular mechanism for the phase separation is suggested.
-
Observation of dynamic crossover and dynamic heterogeneity in hydration water confined in aged cement paste
Journal of Physics: Condensed Matter 20, 502101
-
Studies of phononlike low energy excitations of protein molecules by inelastic x-ray scattering
Phys. Rev. Lett. 101, 135501
Molecular dynamics simulations and neutron scattering experiments have shown that many hydrated globular proteins exhibit a universal dynamic transition at TD=220 K, below which the biological activity of a protein sharply diminishes. We studied the phononlike low-energy excitations of two structurally very different proteins, lysozyme and bovine serum albumin, using inelastic x-ray scattering above and below TD. We found that the excitation energies of the high-Q phonons show a marked…
Molecular dynamics simulations and neutron scattering experiments have shown that many hydrated globular proteins exhibit a universal dynamic transition at TD=220 K, below which the biological activity of a protein sharply diminishes. We studied the phononlike low-energy excitations of two structurally very different proteins, lysozyme and bovine serum albumin, using inelastic x-ray scattering above and below TD. We found that the excitation energies of the high-Q phonons show a marked softening above TD. This suggests that the large amplitude motions of wavelengths corresponding to this specific Q range are intimately correlated with the increase of biological activities of the proteins.
-
The Low-Temperature Dynamic Crossover Phenomenon in Protein Hydration Water: Simulations vs Experiments
J. Phys. Chem. B 112, 1571-1575
A super-Arrhenius-to-Arrhenius dynamic crossover phenomenon has been observed in the translational α-relaxation time and in the inverse of the self-diffusion constant both experimentally and by simulations for lysozyme hydration water in the temperature range of TL = 223 ± 2 K. MD simulations are based on a realistic hydrated powder model, which uses the TIP4P-Ew rigid molecular model for the hydration water. The convergence of neutron scattering, nuclear magnetic resonance and molecular…
A super-Arrhenius-to-Arrhenius dynamic crossover phenomenon has been observed in the translational α-relaxation time and in the inverse of the self-diffusion constant both experimentally and by simulations for lysozyme hydration water in the temperature range of TL = 223 ± 2 K. MD simulations are based on a realistic hydrated powder model, which uses the TIP4P-Ew rigid molecular model for the hydration water. The convergence of neutron scattering, nuclear magnetic resonance and molecular dynamics simulations supports the interpretation that this crossover is a result of the gradual evolution of the structure of hydration water from a high-density liquid to a low-density liquid form upon crossing of the Widom line above the possible liquid−liquid critical point of water.
-
Insights into Hofmeister Mechanism: Anion and Degassing Effects on the Cloud Point of Dioctanoylphosphatidylcholine/Water Systems
J. Phys. Chem. B 111, 589-597
-
Organogels from a Vitamin C-based Surfactant
J. Phys. Chem. B 111, 11714-11721
A new double chained surfactant, 2-octyl-dodecanoyl-6-O-ascorbic acid (8ASC10), with a l-ascorbic acid unit as the polar headgroup was synthesized for the first time. The behavior of the compound in the dry solid state has been characterized through DSC, XRD, and SAXS measurements. The surfactant forms stable viscous organogels in the presence of suitable organic solvents and also water-induced organogels upon addition of water to the organogel. These mixtures show shear-thinning properties and…
A new double chained surfactant, 2-octyl-dodecanoyl-6-O-ascorbic acid (8ASC10), with a l-ascorbic acid unit as the polar headgroup was synthesized for the first time. The behavior of the compound in the dry solid state has been characterized through DSC, XRD, and SAXS measurements. The surfactant forms stable viscous organogels in the presence of suitable organic solvents and also water-induced organogels upon addition of water to the organogel. These mixtures show shear-thinning properties and are birefringent. The behavior and properties of the organogels have been studied through rheology, DSC, and SAXS experiments. The organogels possess the same antioxidant properties of the original l-ascorbic acid ring and can be used to solubilize and protect valuable organic molecules.
-
Specific Anion Effects on the Optical Rotation of alpha-Amino Acids
J. Phys. Chem. B 111, 10510-10519
Patents
-
Machine-learned classification of network traffic
Issued WO2024116129A1
A method for classifying network traffic includes accessing processed network data that includes multiple groups. Each group includes multiple URL data objects and is associated with an entity. The method includes generating a dynamic intent score for each group by generating a comparison value for each URL data object within a group, selecting highest comparison values for the URL data objects within the group, generating the dynamic intent score by averaging the selected highest comparison…
A method for classifying network traffic includes accessing processed network data that includes multiple groups. Each group includes multiple URL data objects and is associated with an entity. The method includes generating a dynamic intent score for each group by generating a comparison value for each URL data object within a group, selecting highest comparison values for the URL data objects within the group, generating the dynamic intent score by averaging the selected highest comparison values, and ranking the groups according to their respective dynamic intent scores. The comparison value for each URL data object is generated by scraping a webpage associated with a URL data object to generate a first scraped text data object, creating web embeddings by providing the scraped text data object to a machine learning module, and generating a comparison value by comparing the web embeddings with reference embeddings.
Other inventorsSee patent -
Methods and systems for automated generation of personalized messages
Issued US WO2018209254A1
A system includes a set of crawlers that find and retrieve documents from an information network, an information extraction system, a knowledge graph storing nodes and edges that connect them, wherein each node represents a respective entity of a corresponding entity type of a plurality of entity types, and wherein the knowledge graph further stores event data relating to events detected by the information extraction system, a machine learning system that trains models that are used in…
A system includes a set of crawlers that find and retrieve documents from an information network, an information extraction system, a knowledge graph storing nodes and edges that connect them, wherein each node represents a respective entity of a corresponding entity type of a plurality of entity types, and wherein the knowledge graph further stores event data relating to events detected by the information extraction system, a machine learning system that trains models that are used in connection with at least one of entity extraction, event extraction, recipient identification, and content generation, a lead scoring system that scores the relevance of information to an individual and references information in the knowledge graph, and a content generation system that generates content of a personalized message to a recipient who is an individual for which the lead scoring system has determined a threshold level of relevance.
-
Method and Apparatus for Dynamic Information Visualization
Issued US 8,683,389
A method for organizing, visualizing, and navigating data objects within a dataset is presented, including network visualization alongside a detailed information display. The method includes interactivity with the visualization through network panning, zooming, and node clicking. The network nodes represent the data objects. Node proximity and links represent similarity or other relatedness measures between respective data objects. The network can be displayed in full to illustrate large scale…
A method for organizing, visualizing, and navigating data objects within a dataset is presented, including network visualization alongside a detailed information display. The method includes interactivity with the visualization through network panning, zooming, and node clicking. The network nodes represent the data objects. Node proximity and links represent similarity or other relatedness measures between respective data objects. The network can be displayed in full to illustrate large scale properties. Clicking the nodes provides further data object information of the highlighted node, which is displayed alongside the network.
Other inventorsSee patent -
Multi-service business platform system having entity resolution systems and methods
US20210357378A1
The disclosure is directed to various ways of improving the functioning of computer systems, information networks, data stores, search engine systems and methods, and other advantages. Among other things, provided herein are methods, systems, components, processes, modules, blocks, circuits, sub-systems, articles, and other elements (collectively referred to in some cases as the “platform” or the “system”) that collectively enable, in one or more datastores (e.g., where each datastore may…
The disclosure is directed to various ways of improving the functioning of computer systems, information networks, data stores, search engine systems and methods, and other advantages. Among other things, provided herein are methods, systems, components, processes, modules, blocks, circuits, sub-systems, articles, and other elements (collectively referred to in some cases as the “platform” or the “system”) that collectively enable, in one or more datastores (e.g., where each datastore may include one or more databases) and systems, the creation, development, maintenance, and use of a set of custom objects for use in a wide range of activities, including sales activities, marketing activities, service activities, content development activities, and others, as well as improved methods and systems for sales, marketing and services that make use of such entity resolution systems and methods as well as custom objects.
Other inventorsSee patent
Languages
-
Italian
Native or bilingual proficiency
-
English
Native or bilingual proficiency
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More