Everyone eats. Many people cook. But few are aware of the unique creative challenges facing profe... more Everyone eats. Many people cook. But few are aware of the unique creative challenges facing professional product developers in flavor and food product companies. The IBM Research AI for Product Composition is a system to assist product developers with these real-life creativity challenges for which no AI-solution existed. The system for compositional product development is in use at McCormick & Company, a global leader in flavor. We present the first set of products targeted for retail sale that has been enabled by the system. The three seasoning mixes for one dish meals are planned to be on retail shelves mid-2019.
There is an art and a craft to making an appealing fine fragrance that takes perfumers many years... more There is an art and a craft to making an appealing fine fragrance that takes perfumers many years to master. The IBM Research AI for Product Composition system was created to assist perfumers to design novel fragrances that are appealing, well balanced and technically sound. The system has been used by Master Perfumers at Symrise to develop 2 fragrances for Brazilian millennials. This creative submission presents the two fragrances, one designed for men and the other for women.
Concrete is the most widely used engineered material in the world with more than 10 billion tons ... more Concrete is the most widely used engineered material in the world with more than 10 billion tons produced annually. Unfortunately, with that scale comes a significant burden in terms of energy, water, and release of greenhouse gases and other pollutants. As such, there is interest in creating concrete formulas that minimize this environmental burden, while satisfying engineering performance requirements. Recent advances in artificial intelligence have enabled machines to generate highly plausible artifacts, such as images of realistic looking faces. Semi-supervised generative models allow generation of artifacts with specific, desired characteristics. In this work, we use Conditional Variational Autoencoders (CVAE), a type of semi-supervised generative model, to discover concrete formulas with desired properties. Our model is trained using open data from the UCI Machine Learning Repository joined with environmental impact data computed using a web-based tool. We demonstrate CVAEs ca...
2014 IEEE 64th Electronic Components and Technology Conference (ECTC), 2014
In this paper, a novel assembly and packaging approach is proposed for 3D/2.5D chip stacks based ... more In this paper, a novel assembly and packaging approach is proposed for 3D/2.5D chip stacks based on bumped substrates. The thinned chips are stacked using thermal compression bonding with “flat” metallization to reduce assembly complexity associated with conventional controlled-collapse-chip-connection (C4) solder bumps. Meanwhile, the laminate substrates are bumped with C4s using injected molten solder (IMS) processes. The pre-stacked chips are then assembled and packaged on the bumped laminates successfully.
2013 IEEE 63rd Electronic Components and Technology Conference, 2013
ABSTRACT In this work, Pb-free microbumps at 50μm pitch with NiFe-based Ball-limiting-metallurgy ... more ABSTRACT In this work, Pb-free microbumps at 50μm pitch with NiFe-based Ball-limiting-metallurgy (BLM) are fabricated and tested. Detailed microstructural analysis has been performed, which shows a uniform thin layer (~0.2μm) of FeSn2 Intermetallic Compound (IMC) is formed between the Pb-free solder and NiFe BLM after the first reflow. In comparison, the NiCuSn IMC can grow more than 2μm in microbumps with the conventional Ni BLM or solder-caped Cu pillar after the first reflow. An excessive lateral “thermal-undercut” has been discovered in NiFe BLM structures during solder reflows due to good edge wettability of NiFe. A dual-layer BLM structure is proposed and demonstrated to mitigate the “thermal-undercut”. Moreover, addition of NiFe layer on a micro-Cu pillar structures have been demonstrated and characterized.
Electronic Components and Technology Conference, 2010
This paper reports a thin wafer handling technology that is compatible to CMOS processing conditi... more This paper reports a thin wafer handling technology that is compatible to CMOS processing conditions to enable 3D integration and assembly with high throughput at low cost. Using pulsed ultraviolet (UV) radiation from excimer lasers, device wafers as thin as 50µm can be released from the temporary mechanical handler wafer in less than 1min. Bonding, adhesive, debonding and post debond
This paper reports a thin wafer handling technology that is compatible to CMOS processing conditi... more This paper reports a thin wafer handling technology that is compatible to CMOS processing conditions to enable 3D integration and assembly with high throughput at low cost. Using pulsed ultraviolet (UV) radiation from excimer lasers, device wafers as thin as 50µm can be ...
We study the refractive index sensitive transmission of a 3D plasmonic crystal that consists of a... more We study the refractive index sensitive transmission of a 3D plasmonic crystal that consists of a square array of subwavelength cylindrical nanowells in a polymer conformally coated with a gold film. Using extensive 3D finite-difference time-domain simulations, we investigate ...
This article presents experimental and computational studies of a phase shifting photolithographi... more This article presents experimental and computational studies of a phase shifting photolithographic technique that uses binary elastomeric phase masks in conformal contact with layers of photoresist. The work incorporates optimized masks formed by casting and curing ...
This work provides plasmonic crystal platforms for quantitative imaging mode biosensing and multi... more This work provides plasmonic crystal platforms for quantitative imaging mode biosensing and multispectral immunoassays, establishing and validating both the optical and equilibrium bases for their operation. We investigated the distance-dependent refractive index sensitivity of ...
AI for fragrance design builds on IBM’s work on Chef Watson to create a system for designing new ... more AI for fragrance design builds on IBM’s work on Chef Watson to create a system for designing new fragrances. Our system generates alternative fragrance formulations and selects the formulation that best optimizes design objectives. We use machine learning to predict both the technical performance of candidate fragrances (shelf stability and skin irritation) as well as the human response (pleasantness and gender appropriateness). We identify if a fragrance is novel by comparing its notes to a large set of commercially available fragrances. The end goal is to use artificial intelligence to design award winning new fragrances.
Everyone eats. Many people cook. But few are aware of the unique creative challenges facing profe... more Everyone eats. Many people cook. But few are aware of the unique creative challenges facing professional product developers in flavor and food product companies. The IBM Research AI for Product Composition is a system to assist product developers with these real-life creativity challenges for which no AI-solution existed. The system for compositional product development is in use at McCormick & Company, a global leader in flavor. We present the first set of products targeted for retail sale that has been enabled by the system. The three seasoning mixes for one dish meals are planned to be on retail shelves mid-2019.
There is an art and a craft to making an appealing fine fragrance that takes perfumers many years... more There is an art and a craft to making an appealing fine fragrance that takes perfumers many years to master. The IBM Research AI for Product Composition system was created to assist perfumers to design novel fragrances that are appealing, well balanced and technically sound. The system has been used by Master Perfumers at Symrise to develop 2 fragrances for Brazilian millennials. This creative submission presents the two fragrances, one designed for men and the other for women.
Concrete is the most widely used engineered material in the world with more than 10 billion tons ... more Concrete is the most widely used engineered material in the world with more than 10 billion tons produced annually. Unfortunately, with that scale comes a significant burden in terms of energy, water, and release of greenhouse gases and other pollutants. As such, there is interest in creating concrete formulas that minimize this environmental burden, while satisfying engineering performance requirements. Recent advances in artificial intelligence have enabled machines to generate highly plausible artifacts, such as images of realistic looking faces. Semi-supervised generative models allow generation of artifacts with specific, desired characteristics. In this work, we use Conditional Variational Autoencoders (CVAE), a type of semi-supervised generative model, to discover concrete formulas with desired properties. Our model is trained using open data from the UCI Machine Learning Repository joined with environmental impact data computed using a web-based tool. We demonstrate CVAEs ca...
2014 IEEE 64th Electronic Components and Technology Conference (ECTC), 2014
In this paper, a novel assembly and packaging approach is proposed for 3D/2.5D chip stacks based ... more In this paper, a novel assembly and packaging approach is proposed for 3D/2.5D chip stacks based on bumped substrates. The thinned chips are stacked using thermal compression bonding with “flat” metallization to reduce assembly complexity associated with conventional controlled-collapse-chip-connection (C4) solder bumps. Meanwhile, the laminate substrates are bumped with C4s using injected molten solder (IMS) processes. The pre-stacked chips are then assembled and packaged on the bumped laminates successfully.
2013 IEEE 63rd Electronic Components and Technology Conference, 2013
ABSTRACT In this work, Pb-free microbumps at 50μm pitch with NiFe-based Ball-limiting-metallurgy ... more ABSTRACT In this work, Pb-free microbumps at 50μm pitch with NiFe-based Ball-limiting-metallurgy (BLM) are fabricated and tested. Detailed microstructural analysis has been performed, which shows a uniform thin layer (~0.2μm) of FeSn2 Intermetallic Compound (IMC) is formed between the Pb-free solder and NiFe BLM after the first reflow. In comparison, the NiCuSn IMC can grow more than 2μm in microbumps with the conventional Ni BLM or solder-caped Cu pillar after the first reflow. An excessive lateral “thermal-undercut” has been discovered in NiFe BLM structures during solder reflows due to good edge wettability of NiFe. A dual-layer BLM structure is proposed and demonstrated to mitigate the “thermal-undercut”. Moreover, addition of NiFe layer on a micro-Cu pillar structures have been demonstrated and characterized.
Electronic Components and Technology Conference, 2010
This paper reports a thin wafer handling technology that is compatible to CMOS processing conditi... more This paper reports a thin wafer handling technology that is compatible to CMOS processing conditions to enable 3D integration and assembly with high throughput at low cost. Using pulsed ultraviolet (UV) radiation from excimer lasers, device wafers as thin as 50µm can be released from the temporary mechanical handler wafer in less than 1min. Bonding, adhesive, debonding and post debond
This paper reports a thin wafer handling technology that is compatible to CMOS processing conditi... more This paper reports a thin wafer handling technology that is compatible to CMOS processing conditions to enable 3D integration and assembly with high throughput at low cost. Using pulsed ultraviolet (UV) radiation from excimer lasers, device wafers as thin as 50µm can be ...
We study the refractive index sensitive transmission of a 3D plasmonic crystal that consists of a... more We study the refractive index sensitive transmission of a 3D plasmonic crystal that consists of a square array of subwavelength cylindrical nanowells in a polymer conformally coated with a gold film. Using extensive 3D finite-difference time-domain simulations, we investigate ...
This article presents experimental and computational studies of a phase shifting photolithographi... more This article presents experimental and computational studies of a phase shifting photolithographic technique that uses binary elastomeric phase masks in conformal contact with layers of photoresist. The work incorporates optimized masks formed by casting and curing ...
This work provides plasmonic crystal platforms for quantitative imaging mode biosensing and multi... more This work provides plasmonic crystal platforms for quantitative imaging mode biosensing and multispectral immunoassays, establishing and validating both the optical and equilibrium bases for their operation. We investigated the distance-dependent refractive index sensitivity of ...
AI for fragrance design builds on IBM’s work on Chef Watson to create a system for designing new ... more AI for fragrance design builds on IBM’s work on Chef Watson to create a system for designing new fragrances. Our system generates alternative fragrance formulations and selects the formulation that best optimizes design objectives. We use machine learning to predict both the technical performance of candidate fragrances (shelf stability and skin irritation) as well as the human response (pleasantness and gender appropriateness). We identify if a fragrance is novel by comparing its notes to a large set of commercially available fragrances. The end goal is to use artificial intelligence to design award winning new fragrances.
Uploads
Papers